JPS61105174A - Facsimile equipment - Google Patents

Abstract

PURPOSE:To obtain a facsimile unit which can efficiently read and encode by using a program control means such as microprocessor for compressing coding and controlling a reading part. CONSTITUTION:When a reading starting signal 18-a is sent out from an AND gate 18, reading part starts reading one line and sends binary picture signal 9-b to line buffer 10 in synchronism with a writing clock 9-a from reading part 9. An F/F-13 becomes OFF when the one line reading operation is started, and when the reading operation is completed, i.e. when all information of one line of 2048 bits is written in the line buffer 10, it becomes ON. A controller 21 controls the sequence of the entire facsimile equipment and also makes signal conversion for the binary picture signals stored in the line buffer 10.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention provides a document for transmission (hereinafter referred to as a transmission source).
The present invention relates to a Fakunmi IJ system that removes the redundancy of a binary image signal and transmits a band-compressed signal. In general, gakushimi') devices are provided as image transmission equipment, but looking at their usage, they are used to transmit characters, symbols, etc. (hereinafter referred to as characters) written on documents, slips, etc. There are many grooves that are used for C. In such a case, it seems that information such as characters and symbols is being transmitted rather than the image itself. For example, both the bold ``A'' and the thin ``A'' represent 1A1, so when transmitting the ``hiko'' character, it is only necessary to convey the meaning of the ``hiko'' character to the receiving side. Generally, the binary double-image signal obtained from the transmitted original includes the line width signal representing the a-line and bold text. Therefore, sufficient time is required to transmit the line width signal.
Ru. Therefore, the above! If the s@ signal is deleted, the transmission time of the MA binary image signal can be shortened. Therefore, when transmitting a manuscript that mainly consists of text information such as the above-mentioned documents and slips, the binary 11ii obtained from the transmitted manuscript
Rather than faithfully transmitting image signals over time, it is better to quickly transmit the meaning of the text information by sending only brief information about the location and length of each part of the text information. I can say that. (The position and length of each line segment determines the character information in the previous 1e, and by adding an arbitrary width to the character information, it is possible to make bold ink or thin ink.) Then, in reception, by inserting a signal approximately equal to the deleted line width signal at a predetermined position, it is possible to obtain approximately the same received image of the transmitted original. The facsimile machine of the present invention is designed to take into account the above point 1, and the facsimile machine of the present invention is designed to take into account the 2 points obtained by scanning the transmission document.
I! From the II image signal, -f, part or all of the line width signal (black run A to be described later) in the character information is deleted. Next, a conventionally used compression process (for example, a one-length encoding process using a Wyle encoding code, which will be described later) is performed. Therefore, since the compression effect is higher than that of the conventional method, the transmission time of the two-fold image No. 1g is further shortened. Next, when the transmission signal 1 is received at the receiving side, a line width signal having an arbitrary width is added to the position of the line width signal from which m16ti has been deleted, and a received image is created.

The present invention will be explained below with reference to the drawings.

@1 figure fal #'i, Transmission 1 that can be applied to the 7mm device of the present invention! Horizontal scanning method, ・Yan 1 figure et al.) ~ (J), compression method of binary image signal obtained by previous 4e scanning, etc. II! Following I, I wrote a self-description. It is something.

1st [i! ! 1 shown in 1fal is an example of a cave manuscript mainly consisting of characters, and 2 indicates a scan @ group.

Scanning lines 2-+, 2-2.2- constituting the scanning line group 2
3, first scans from the left end to the right end along the scan @2-IK, and then scans from the left end to the right end along the scan line 2-2. Furthermore, scan J12-5 and subsequent scan lines (not shown) are also scanned in the same manner as described above.
, the binary image signal of FIG. 1 tbl is the scan #i! 2-
1tC119 represents the signal when scanning, and the signal of the black background part 1-11 etc. is set to black level 3, and the white background part 1
-2 signal is expressed with a white level of 6.

First, when scanning starts from the left end along the scanning line 2-1, the black runs 5a, 3 belonging to the black level 3 are found in the portion of the letter A.
A' is obtained, and the following is an opening and the letter B? Run 3b is
The black run 3C in the letter C is the black run 3d, 3d' in the letter C,
The cicada run 3e is the letter E, the black run Sf is the letter P'', and the letter G
If you use the letter H, you will get a brave run of 5g, and if you use the letter H, you will get a black run of 5h and 5W. In the above-mentioned book, there are white V per VC, % white run, and from the note, white run d-1, 4-
2,......Rule 12. Said heat run 5a-s
kI, day run 4-1 to 4-12, each 7) information amount (bits and symbols are from Table 1).

Table 1 Therefore, since the total amount of information is the sum of the black run bits and the white run bits, it is 140 bits in total. The first @ tel is the abbreviation shown in Table 1 [black run is represented by B, white run is represented by W. The number written after B or W indicates the amount of information (bits). Accordingly, the binary image signals are visually arranged according to the scanning sequence (from left to right) shown in FIG. 1('b).

FIG. 1+a is a transmission code that has been subjected to the compression process (one-dimensional run-length encoding process using Wele's encoding code) based on the abbreviations shown in FIG. 1(C). Here, the one-dimensional run-length encoding process means that the continuous length of the black run and white run in the binary image signal (hereinafter referred to as run length) is encoded and expressed in a binary scale. It is a band compression method that reduces the number of bits, and W
The yle encoding code refers to encoding of the binary image signal by the method shown in FIG. (Other encoding methods include VcHu fman, Golor
Although nb codes and the like have been proposed, they are omitted here. ) The +10110 written at the upper left corner of FIG. 1 1dl corresponds to Wt5 written at the upper left corner of FIG.
Continuing from W15 (corresponds to Bx), and the following corresponds in the same way.Therefore, the code +10110.010 of 1dl in FIG. 1 is 1d in FIG.
The abbreviations W15, B3, . . . of l are encoded, respectively. Here, the Wyle encoding code will be explained. In Wyle's encoding code shown in Figure 2, the encoding code is an address (a code that determines the code length and bias) followed by a remitter (a code that determines the run length). . For example, since the black run 5a (abbreviation B3) is 3 bits, the K used to encode it is the "white or black run length" column.
If you look at the right side of "3", the address is "0" and the reminder is [+oJ, so it becomes l"otoJ. Similarly, if you encode 10 pits, you will get i+oo'01.15 bits (for example, the above abbreviation W15) yk When encoded, it becomes 1101
10 (omitted in FIG. 2). In this way, the abbreviation of tel (binary image signal) in Figure 1 is 1dl in Figure 1.
transmission code. (The "Wyle encoding code" referred to in the present invention is the "white or black run length" 1025 to 1152, the Margin part, and one "O" at the right end of the "address". In addition, the amount of information of the transmission code converted by the Wyle encoding code is O or 1.
The question is how many there are. Therefore, @1 figure 1 dlK
The total amount of information in @1 is 91 bits, which is 1 in Figure 1cJ.
If you compare it with 40 bits, you can see that it is compressed. −
The film pressure compression ratio is expressed as the total number of pits for one line divided by the total number of bits after compression processing, so the compression ratio is 140÷9.
1=154.

In addition, the first block (9110j10 shown in the upper left) is always regarded as a white run, and the white run and black run are alternately represented, so the transmission code has a white There is no need for an identification signal to distinguish between runs and black runs.Therefore, if the transmission code shown in Figure 1 (mountain) is changed to an electrical signal and transmitted, the receiving side will first see Figure 1 1cl.
It is determined how many bits of white run or black / each code consists of (Figure 1, 1cl), and then a binary image signal is created by the above determination [@1, Figure 5], and the above 2 Based on the value image signal, the reproduced image information [@1, which is equal to 2 in FIG. 1a3, is obtained.

Here, for convenience, we will call the conversion of the phrase abbreviation (binary ij signal) in Figure 1 to the transmission code of Figure 1 1cl. (This belongs to the conventional code conversion. )
.

@Figure 1 (The character "al" is made up of a combination K of line segments having a constant width, so the bit of the black run changes depending on the direction (inclination) K of the line segment. Therefore, The black runs include a signal consisting of a relatively short black run (referred to as black run A) to represent a part facing vertically (up and down in the figure) or diagonally, and a line segment facing in the horizontal direction (left and right in the figure). can be divided into two types: a signal consisting of a relatively long black run (referred to as black run B).

The value K of the black run A and the black run B>, and the above @(M
AX) is 9 bits, the lower value (MIN) is 5 bits, black run A is a black run in the range of less than 9 bits and 3 bits or more, and black run B is a black run other than the above (less than 3 bits or 9 bits). above), then the abbreviations in Figure 1 1cl K are BS, BA is black run A, abbreviation B9 is black run B, and so on.

@Figure 1 (- indicates that only the information on black run A is removed from all the information in Figure 1 1cl (consisting of black runs and white runs), and there are consecutive white runs (for example, W15 and W2). It turns out that it was.

@1 figure) encodes the abbreviation of @1 figure tel according to Wyle's encoding code, and furthermore, 1 At the beginning of each code, an identification bit of 0 representing a white run or 1 representing a black run B (marked with *) is encoded. It is a transmission code to which a 0 or 1) is added.

The identification bit was added because it became necessary to distinguish between the white run and the black run B by selectively removing only the black run. Therefore, Fig. 1 if) K
The amount of information of the transmission code in
It becomes 4 bits. The next transmission code (see Figure 1) is changed into an electrical signal, transmitted, and received by the receiving side. Based on the received signal, it is first determined whether the code block to which the identification bit is attached (for example, ff110+10) represents the run length of the white run or the run length of the black run B.

Next, the code is decoded to determine the run length of each run [
Figure 1 (from 0 to Figure 1 + e+) and 9 consecutive 1 white runs, a black run between the consecutive white runs (a black run inserted arbitrarily on the receiving side, the black run Insert (replaces N.Here it is 3 bits) [Figure 1 1cl
to Figure 1]. The black run to be inserted has an appropriate run length determined in advance (corresponding to the width of the line segment in the received recorded image, and considering the line width of a typical original, approximately 15 f).
i), the black run A before transmission becomes black yN on the receiving side. Specifically, in Figure 1 (black run in the binary image signal of phrase and lcl is 3 a = 5 c = 3 d = S d' before transmission)
= 3 f = 5 g = 5h=sh'=Bs, 5
pl=se=Bao relationship [Figure 1 1d] [However, since it is replaced by black run (=B S ) on the receiving side, Sa=3g'=5c=Sd=5d'=36=5f=
If Sg = 3b = 5W = 85 and the transmission code shown in Figure 1 (converted to an abbreviation such as φ. Therefore, Figure 1) is changed to an electrical signal and transmitted, the receiving side will first receive the 1
Figure 1 (determine how many bits of white run or black run B each code of 0 consists of [Figure 1 1e)], then 1
Insert the black run N between consecutive nine white runs in the discriminated nine signals (Figure 1), create a binary image signal from this [@ Figure 1, open can K is approximately equal], and then generate the binary image. Based on the signal, a reproduced image (approximately the same as 1al in FIG. 1) is obtained. Here, the abbreviation (2
＠Image signal) is changed to the abbreviation of 彬) in Figure 1 and:, t
Fig. 41 (conversion to fl transmission code will be referred to as signal conversion B for convenience. (This belongs to the code conversion adopted in the present invention). The signal conversion B converts the bits of black run A. Since the binary image signals reproduced on the receiving side are simply aligned, the binary image signal reproduced on the receiving side is shortened or expanded in the scanning direction (left and right) depending on how the black runs are selected.

[Figure 1 1al is 140 bits, whereas Figure 1 (- is 158 bits]).Improved this point,
In order to prevent the bits in the scanning direction (left and right) from changing, K is the black run to be extracted, rather than simply extracting the black run λ from the abbreviation shown in Figure 1 to create an abbreviation such as 11th ffi te+. It is sufficient to adjust the difference between black run A and black run in advance using the white runs before and after the input. Specifically, since black runs 5i and 5e in FIG. 1c are both 4 bits (Bm), one black bit is subtracted to make them the same bits as the black runs. (B3). Next, the one black bit is regarded as a white bit and is added to the white runs A-2 and A-7 before (or after) the black runs Ag' and 5e. (Figure 1 1al is obtained), and black run A
By removing , we obtain the abbreviation in Figure 1. Figure 1 1alf
i is a transmission code in which the abbreviation in FIG.

Therefore, the information amount of the transmission code in this case is 74 bits. If the transmission code in Figure 1 1al is changed to an electrical signal and transmitted, the receiving side first determines how many bits of white run or black run B each code in Figure 1 (i) consists of. [Figure 1 (5), Next (, insert the black run between the consecutive white spots 7 in the discriminated signal - Figure 1 i)) From this, create a binary image signal [Figure 1 ( Based on the binary image signal, the epiphyte image ['@1 (Fig.
alK approximately equal].

Here, in Fig. 1 (abbreviation of cl (binary Im signal)
For convenience, the conversion to the +00 transmission code will be referred to as signal conversion C (this belongs to the code conversion adopted in the present invention). According to signal conversion C, the total number of bits in -scanning is 1A0 bits, which is the same as the abbreviation in FIG. There is no.

The abbreviation (binary picture signal) in Fig. 1 1c) is 140 bits,
@Figure 1 (The transmission codes for fl and +1+ in Figure 1 are both K”7
Since it is 4 bits, the compression ratio is 140÷74=18
9 and Nasho, when compared with the compression rate (tSa) by K for the run-length encoding process, it is approximately 2λ7% ((154-t
) x 1o O by 22.7%]. As described above, the facsimile apparatus of the present invention using the 1-predistribution signal conversion B or the signal conversion C can further increase the compression rate compared to the conventional run-length encoding processing (the signal conversion A). Next, a detailed explanation will be given of the facsimile apparatus of the present invention.

In @AR, the drive section 5 is composed of a pulse motor and its canter drive circuit, and feeds the document 1 by the scan line group 201 runs in response to a document movement signal 19-aK, which will be described later. The original to be transmitted 1 is held and conveyed by the paper feed roller 6 driven by the drive section 5 and the suppression rollers 6' and K provided opposite to the feed roller 6.

Images such as characters 7 written on the transmission document 1 are read by an optical system 8 (for example, COD).

(composed of an image sensor such as a photodiode array and its driving circuit) forms an image. In the reading section 9,
Scanning for one line is started from a reading/shaping start signal 18-aK, which will be described later. When the image signal for one line is read by the scanning and converted into a binary image signal as shown in Figure 1, the binary image signal is converted to a write clock 9-a.
, OR gate 11, and output 11-1, the signal is written into line buffer 10 via signal line 9-b. At this time, the write clock? -a is counted by a counter 12°. 1 line full clean latwo 20
If the aa bit is used, the counter 12 can be configured to sweep out the carry signal 12-a when it counts 204B. The carry signal 12-
a is buffer full 7 lag fritz ycxyy
' (hereinafter referred to as f) is input to the 130 set terminal. Further, the reset terminal of F/F'13 has 0 or more configuration 1(
k, read start signal 18- from AND gate 1B
When a is sent out, the reading section starts reading one line, and the write clock 9- from the reading section 9 is sent.
a Binary image signal 9-b is sent to line buffer 1° in synchronization with K. P/F'-13#-1: When the reading operation of one line starts, it becomes the 1cOPF state, and when the reading second operation ends, that is, the line buffer 10
When all the information of one line of 2018 bits is written, it becomes ON state. The line buffer 1° is controlled by the reading speed of the reading section 9 (which is approximately constant regardless of the content of the binary image signal) and the encoding processing speed of the controller 21 (which varies greatly depending on the content of the binary image signal). This buffer is provided to smooth the transmission of the image signal by buffering the speed difference between It is a well-known fact that efficiency can be further improved by using /buffers alternately.

The controller 21 controls the entire sequence of the facsimile machine, and also performs the signal conversion B or C described in the explanation of FIG. 1 on the 2#L image signal stored in the line buffer 10.

The controller may be realized by a microcomputer or a hardwired control logic circuit.

Note that since the controller 21 is the most important part in the facsimile apparatus of the present invention, it will be described in detail later along with the decoder 20.

The transmission code created by the controller 21 is transmitted to the First In First
The F'IP (Jf: ++) is stored in the memory 24 (hereinafter referred to as FIK). This is provided to buffer the speed difference with the transmission speed of the transmission code. Regarding the characteristics of PIF & memory, for example, Fajrchild of the United States
Since it is well known and is described in detail in the MOS LSI3351 catalog of the company's product, a detailed explanation will be omitted here. The controller 21 inputs the PI input in the FiFo memory 24! FIRDY sense signal 2 from child (hereinafter referred to as FYRJ)
m-a.

The AND gate 22 is passed to the friend signal 22-af:v4.

At this time, if the PIF□ memory is full, FI□1)
Y is OF'? , if there is room to store data in the FIFO memory, 1; 'IRDY is turned ON. Therefore, the control cooler 21 writes the transmission code to the FIFO memory 24 via the data bus 21-C when FIRDY ON. [At this time, the write clock is shift-in signal 2
3-a, the FIFO speed transmission code is sent from the terminal ou to the parallel-in-serial-out shift register 25 via the signal line 24-b. The shift register 25 outputs the 8-pitono transmission code (lm2) inputted manually through the shift register 24-b.
5-a to the modem 28: The modem 2B sequentially transmits the transmission codes from the shift register 25 in synchronization with the transmission clock 28-bK. The transmission clock 28 is simultaneously input to the 8m counter 21 and counted. The 8-way kawanta 27 sends out a carry signal 27-a every time it counts eight. The carry signal 27-a is input to the LOAD terminal of the shift register 25. With this configuration, every time the Modeno 28 finishes transmitting an 8-bit transmission code, the shift register 2
5, the new 8-bit transmission code is transferred to FIFO memory 2A.
It is output from

In addition, the carry signal 27-ati and the delay circuit 26
Also, the output 26- of the delay circuit 26
ali, FIFO memory 240'/7 act (hereinafter S
(denoted as O) terminal.

With this configuration, after the shift register 25 latches a new 8-bit transmission code, the shift-out signal 26-a is applied to the %PIF (J memory).

The modem 2B consists of a modulation/demodulation circuit and a level adjustment circuit (not shown). Output line 25 from the shift register 25
-a, the transmission code taken in via modem 2
The signal is input to the modulation/demodulation circuit (not shown) at 8 and modulation is performed. A facsimile machine transmits an image signal through a telephone line, and the frequency band that can be transmitted by a telephone line is generally 300Hz7), raaaK)iZ and ntt.
lAru. Shikashi, transmission code 1 from shift register 25
The code contains a combination of c#'i DC output 17(J)fZ, and as it is, the code cannot be transmitted over the telephone line. Therefore, by setting the carrier 11Ft number to a frequency that can be transmitted over a telephone line and modulating the carrier number with the transmission code, it becomes possible to transmit the transmission code over the telephone line. (Carrying wave at - radio AM variation A", F
The principle is roughly the same as MR@. It is also possible to employ KPM modulation and other variations. ) Also, when transmitting the code, since signals are exchanged with the other party (for example, the transmitting side of the transmission code receives a confirmation signal that the transmission code has been received), the modulated transmission signal from the other party Since the sending side of the transmission code receives a confirmation signal indicating that it has received the need to demodulate the transmission code, it may be necessary to demodulate the modulated transmission signal from the other side. The modulation/demodulation circuit (not shown) performs the above-mentioned modulation and demodulation, converts the code from the shift register 25 into a transmittable signal, and demodulates the signal sent from the other party. The output of the modulation/demodulation circuit (not shown) is input to the level adjustment circuit (not shown). The level adjustment circuit (not shown) is capable of controlling A to the telephone line network caused by excessive output of the modulated transmission code, excessive input of the signal sent from the other party (thus causing an adverse effect on the device, and similarly insufficient output). , is provided to prevent troubles caused by insufficient input.In this way, the modem 28
The system modulates the transmission code sent from the shift register 25, demodulates the signal sent from the other party, and adjusts the level of the transmission code, etc., to ensure smooth transmission and reception of signals by the telephone IsK. be. and output line 28-a
The encoded signal is sent from

Next, an internal configuration diagram of the controller 21 is shown in FIG. d. Also, the CPU in the controller is 1nte11080
0 internal configuration diagram is shown in FIG. The controller 21 is a microcomputer 6080.21- manufactured by 1nte1 in the United States.
Mainly d, ROM (31 product names 8316 and below), 2
l-eSR person M (8101) 21-fl 5TATU8
LATCH (8212) 21-1. It is composed of path drivers (11212) 21-gs21-h, 21-t, and clock pulse generators 2l-kK. Also,
IA-a, 15-a, 22-a in Figure 5
is represented by 21-oK in FIG. Regarding the basic operation of the controller and the operation inside the CPU, the 8080 Microcom of 1nte1 Co., Ltd.
puter 8systemsUser's Manua
8080 software is detailed in the literature such as 9, and the 8080 person sse of 1nie1 company.
mbly Language Programming
Since it is shown in detail in the literature such as Manual, detailed explanation will be omitted here. The interface between the controller 21 and the external I10 will be described below.

Address path 21-b is input to decoder 20 to generate each stack selection signal. The relationship between each stack and the decoder output is shown in Table 2 below.

Table 2 As shown in Table 2, each decoder output passes through a gate to generate a control signal for each I10K. For example, in FIG. 3, the original elliptical drive signal -19-al given to the drive unit 5
a. The drive section selection signal 20-C and the controller 21
The W signal 21-a from the AND gate 19 is logically AN
D is created as a result of taking the 6th order.Through the above configuration created by each gate, the information processing for one line proceeds, and furthermore, the document drive signal 19-a is sent from the controller 21 in FIG. is input to the drive section 5, the feed a--2 and the pressure cooler 6' feed the transmitted document 1 by a distance of K1247 in the sub-scanning direction (up or down in the figure) [Fig.
From scanning line 2-1 to scanning line 2-2 shown by aliC] The next line (scanning line 2-2) is read by the reading section 9.

Thereafter, scanning is performed in the same manner, and the plan of the transmitted original 1 is scanned.

The 7a-chart shown in Figure 6 (Canada, a311GcJ) represents the information processing error of the symbol conversions B and C performed in the controller 21, and the ?? end of the scanning line in the 65th car is black. Shows the processing method when .

In addition, the route table shown in Figure 7 represents the processing route when processing is performed according to the flowchart in tIC6 Figure (2), and * is processed by adding the right even route. It represents. Furthermore, FIG. 8 1al is =t n)
ts-9 internal ROMzl-e, Figure 8b) is RA
The internal memory maps of M21-f are shown respectively.

First, the symbols used in FIG. 6(a)K will be explained below.

RL: 11S bit Run length counter 2l-f-i (
(shown in Figure 8) ML: 16 bits, white run length register 2l-f-ri (
(shown in figure M8) BL: 16 bits Black run length register 2l-f-c (
) LCNT: 16-bit line counter 2l-f-ho (-) D: New data memory 2l-d-i (shown in Figure 5) MOD: Old data memory 21d-port (shown in Figure 5) B: Black Represents a bit (=1) W stored in the program in the ROM: Represents a white bit (=Da) MIN: Constant (=3) MAX stored in the program in the ROM: Constant ( =9) BR: a bit operation result memory 2l-f-2 (shown in Figure 8) DIS: @separate bit memory 2l-d-c (shown in Figure 5) Furthermore, "=" in Figure 6) The second symbol represents a subroutine. 1 sub, routine is 6th
Since it is not described in the figure, it will be briefly explained below. #The subroutine is W7, which was explained in Fig. 2 above.
1e encoding method, the data stored in the RL is converted into the WyIe code shown in FIG. 2, and the data is sent to the FIFO memory 24 in 8-bit units. The FIPO memory 24 stores transmission codes in the format shown by +r+ in FIG. Said W
y I e encoding method (detailed explanation is given in Wyl
Is it the encoding method of e itself or Wyl, e's paper (
H,Wyle eta1″Reduced-time
facsimle transmission by
digital codig"IRJi: Trans
, Com-9, P215 (1961-09)), and it is obvious that it can be easily realized by the controller 21 or a device having an equivalent function, so the explanation will be omitted here. In addition, the black run 8 and run length (note 1) omitted in Figure 6 1al are
RL) is a black run that satisfies the condition of MIN × RL ≧ M input X, the operation performed in (Note 2) is an integer maximum, and (Note 3
), the black at both ends of each line is always transmitted, as shown in the example shown at the bottom of FIG. 6(a).

Next, I will explain Fig. 7 and explain the operation of the flowchart in Fig. 6 (al).
The processing will be explained by considering one line of the marked two-color image signal. In FIG. 6 1al and FIG. 7, first, route 29
Start from. Per K, which processes one line,
As an initial setting, WL.

Write BL, BRK plays. Also, write the negative bit number of one line, in this case (-L(S)) to LCNT. Next, execute the K-ad-Hit subroutine (see FIG. 6). Here, the controller 521 is the line buffer selection Shinto 20-
By outputting a, one bit of the line buffer 10 is first taken in as a binary image signal 15-a through the AND gate 15, and the one bit is combined with new data and the new data image v.
-I) (21-d-I) Write K. Next, add +1 to the contents of KLCNT.

This completes the work of the °d subroutine. Next, R
Write 1 to L, then proceed to route 30 and MO the contents of D.
Dzl-(', - after writing in the mouth, on route 31,
Execute M subroutine. After that, MUD (21-d
- 口) and D (2f-d-i) are determined to be equal. In this case, it becomes MOD'qD, so route 3
3, and then it is determined whether MOI)=W.

In this case, KFiYES, so go through route 37 and WL
The contents of KRL (in this case RL=1) are written, and then 1 is written to RL to determine whether LCNT=da or not. Since LCNT=-1t, it is NO and the route is 50.
and return to route 30.

Therefore, the contents of D21-d-1 are written to the MUD (21-d-port) and the W subroutine is executed. After that, it is determined whether MOD=D. Posture is Y
Since it is ES, route 32 is lost, write RLK2 and LCNT=
judge whether it is wrong or not. Since LCNT=-1s, NO and Nasho,
Return to path 31 through path 35 and repeat the above operation, and the sixth data counted from the left end of the binary compensation signal for one line consisting of 16 bits is D(21-d-i)K15. The th data is MOD (z 1-d-mouth)K
When the data is written and RL=a and LCNT=-1o, the determination as to whether 1MUD=D becomes NO, and the process proceeds to path 34 via path 53. Since the determination as to whether the next KMOD=W is NO, the process proceeds to route 36. Here, RL is! Since the condition of (RL-M
IN), and calculate the difference as BR(21-f
-2) Incorporate CIF. Next, it is determined whether KWL=g or not. This determination is made to determine whether the black run whose length is currently 70 runs stored in RL (21-f-i) is the black run at the left end of one line.

Since WL=1 here, the process proceeds to route -0.

Next, it is determined whether LCNT=〆. This judgment is
RL (21-f-port) K The black run length whose run length is memorized is the right end of one line (this is done to determine whether it is a certain black run or not. Here, LCNT==-10, so route 42 and proceed to BL (21-f-no) I
/c96tilF! Including! , WLK (contents of WL) and (B
The contents of R/2) are added and the entire result is written, in this case B'
Since FL=1, the result of integer calculation is rounded down to WL
= 1 + engineering = 1.

Next, it is determined whether BR is an odd number or not. Since BR=1, it is YES and proceeds to route 45 WL (21-f-exit) K1
Write +1=2, then write BRK da path 4
Proceed to step 8. Next, BL8 (21-d-Is) [W(+=
Write WL (21-f-i) to RL (21-f-i).
After writing the contents of f-mouth) tc raw stone] ■ Sub・
Proceed to routine.

[See Figure 6 1cl].

[In the Roroko = 10 Napaku routine, it is determined whether or not Masu RL = /. In this case, since KL=2, the contents of the next KDI8 (21-d-c) (in this case, information f indicating a white run is stored in DI8) are sent to the FIFO memory 24, and then m Statement oRL**! Execute the al subroutine.

This completes the processing of the g subroutine. Then, return to main-70-.

Next, write B (=1) to BL8 (21-d-c), send it to the self-answer t-FIFO memory 24, and then write the above-mentioned RL**! Execute the llt$ subroutine. Then, write (B + 1) to RL (21-f-i), then write KBR (21-f-2) and BL (21-f-c).
After writing each of K, it is determined whether LCNT=Me or not. Since LCNT==10, NO and tL are returned to route 50 via route 51. And route 51-route 32-route 35t10 times < lK ffl, RL=11, I, C
With N'l'=/, MOD-D=/, route 3
Proceed to step 4. Y18 after determining whether KMOD=W or not.
Therefore, the contents of RL (21-f-a) are written to WL (21-f-a)K through path 57, and RL (21-f
-B) Write Kl and judge whether I or CN'l'=Me, and since it is YB8, proceed through route 49 to route 48.

Then repeat the same action as above. This time, WL=11
, BL-〆, only the white run is fiRL-converted and sent by the island 1] nap ψ routine.

Then, the process proceeds to route 52, and since the determination as to whether it is MOD, .

In addition to the processing process for the 1-ray y-minute 02-value image signal that has been explained above, the processing process for the 02-value image signal is shown in Figure 117. Regarding the specific processing operations inside the controller 21, such as writing to memory, reading operations, calculations, size judgment, match judgment, etc., please refer to the two books published by 1ntel.
Since Fl-mK is described in erk Ma+nual, the explanation is omitted here.

In addition, as I have explained more than once in Kontoku 221, 90 PUs can also be used with OCR technology if you use a high-end CPU (one that is as good as 1ntel 808Ω in terms of processing speed and processing speed of 11a). It is possible to perform 6-sided trimming processing using known technology, and the line width of image information can be reduced (
Before performing the signal conversion C), if K) is kept constant, I
! The applicable range of signal conversion C can be further expanded.

As explained above, the factor cutting device of the present invention can significantly shorten the information transmission time by applying appropriate processing to the above-mentioned black numbers. However, if the transmission document is made up of image information, if the above-mentioned key conversions B and C or the above-mentioned thinning processing are performed, the ≦Shinki Kama image on the receiving side will be significantly degraded. For this reason, it is necessary to appropriately select whether or not to apply the signal conversions B and C to a transmission document containing image information in order to prevent image deterioration. If the transmitted document contains image information and it is inappropriate to perform the signal conversion B and Ct, set the threshold value to OK (
MAX=MIN=O) can be set. (The black run is gone and all the white and black runs are transmitted. This will do the signal converter). This increases the transmission time, but improves the image quality.

As described above, the facsimile device of the present invention not only reduces the transmission time for characters, symbols, etc. having a substantially constant line width, but also reduces the transmission time for characters, symbols, etc. having different line widths by performing the thinning process. Also, rapid transmission becomes possible.

In addition, the threshold value between the black run person and black run B (M people x 1 ML
If N) is set to an appropriate value, normal image information can also be transmitted.

[Brief explanation of the drawing]

Figure 1: Figure 1 is a front view of the transmitted manuscript, which mainly consists of text.
Figure 1 (bl is the binary image signal obtained by scanning the transmitted original in Figure 1 (al), Figure 1 (e) is the abbreviation for the binary image signal in Figure 2, Figure 1 (dl is the symbol in Figure 1) (Transmission code compressed based on the abbreviation of C1, tel in FIG. 1 is the abbreviation of signal conversion B applied to the abbreviation of C1, FIG. 1 (fl is based on the abbreviation of FIG. 1(e)) After receiving the transmission code in which an identification code is preliminarily added to the bundle, the transmission code shown in Fig. 1 (- is shown in Fig. 1 (f) is received and converted back to the abbreviation of el in Fig. The abbreviation in Figure 1 (h) is the abbreviation of the term "White Run" plus "Black Run".
FIG. 1(i) is a transmission code in which compression processing is performed based on the abbreviation in FIG. 1(hl) and an identification code is added to the beginning. Happy 1st figure (
j) receives the transmission code of FIG. 1(1) and transmits the transmission code of FIG. 1(h
The abbreviation is changed back to the abbreviation of l, and the abbreviation is changed to K, and the black run is added to the consecutive white runs.
. ω encoding code, FIG. 5 is a circuit configuration diagram of a facsimile apparatus according to the present invention, FIG. 4 is a configuration diagram of a controller, m
s is an internal configuration diagram of the CPU, FIG.
Figure Φ) is a 7CI-chart diagram of RaadBit in Figure 6 Tal, Figure 6 (cl is 7a-f+- of Sen RL in Figure 6 (al), Figure 6 (d)
is an explanatory diagram of the process when both ends of the scanning line are black runs, *71f
Figure 6 (al~(cl)y a -5-q)
FIG. 8(b) is an internal memory map diagram of OM, and FIG. 8(b) is an internal memory map diagram of tiR person M. In the figure, 1 is the transmission source 4115, the drive unit 6 is the feed roller, ◇ is the pressure roller 8 is the optical system, and 9 is lIi! 4L circuit 10 is a line buffer 11 is an OR gate 12 is a counter 13 is a buffer full flag flip-flop 14.15
．． 17.18.19.22.23 is ANL) Game 116
is NOT@@20 is decoder 21 is j7toa
-La 21-d is CPU (8080)21-
C is ROM21-f is 8 M24 is FIFO memory 25 is shift register 26F
i-delay circuit 27 is 81 counter 28 is modem Patent applicant Canon Co., Ltd. 68 Figure (Kuji <b) Procedural amendment stamped) November 5, 1985 1, case display October 4, 1985 Patent application submitted on date (5) No suffix 2, title of invention 3, relationship with the case of the person making the amendment Patent applicant address Address: 3-30-2 Shimomaruko, Ota-ku, Tokyo Address: 146
3-30-25, Shimomaruko, Ota-ku, Tokyo, the entire text of Statement of the Subject of Amendment 6 and Statement of Contents of Amendment shall be amended as attached. Description 1, Title of the Invention Facsimile device 2, Claims Reading means for reading the image of the transmitted manuscript. A program control means having a program memory storing an encoding program for capturing a z-valued image signal obtained by a reading means and sequentially compression-encoding it into a code for transmission, and a control program for controlling the operation of the reading means, a compression code. 1. A facsimile device comprising: a code memory that temporarily stores encoded transmission codes; and a modulation unit that modulates data read from the code memory for transmission. 3. Detailed Description of the Invention The present invention scans a document to be transmitted (hereinafter referred to as a transmission document) and obtains a white or black signal (hereinafter referred to as a binary image signal). This invention relates to a facsimile machine that removes redundancy and transmits a band-compressed signal. <Prior Art> Conventionally, a facsimile machine as disclosed in, for example, Japanese Unexamined Patent Publication No. 49-49520 has been known. Such a device performs run-length encoding on the read image signal,
Furthermore, a series of processing to convert the data into compressed code was performed in hardware. This required a complex circuit. Furthermore, since the control of the reading section is performed by hardware, the synchronization relationship between compression encoding and the control of the reading section becomes complicated. Purpose C> In view of the drawbacks of the prior art as described above, the present invention efficiently executes reading and encoding processes by using program control means such as a microprocessor for compression encoding processing and reading unit control. The purpose is to provide a facsimile machine that can. Embodiment> Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 (1) shows a method of scanning a transmitted document in the facsimile machine of this embodiment, and FIGS. , are described in order. 1 shown in FIG. Scanning line 2-1゜2-2.2-
3 scans first from the left end to the right end along the scanning line 2-1, and then from the left end to the right end along the scanning line 2-2. Furthermore, scanning line 2-3. The subsequent scanning lines (not shown) are also scanned in the same manner as described above. The binary image signal in FIG. 1(b) represents a signal obtained when scanning is performed along the scanning line 2-1. Signal of black background part 1-1 etc. at black level 3. The signal of the white portion 1-2 is expressed at its own level 4. First, when scanning starts from the left end along the scanning line 2-1, the black runs 3a, 3 belonging to the black level 3 are found in the portion of the letter A.
ai is obtained, and in the same way, black run 3b with letter B is
In the letter C, the black run 3c is a letter, and the black run 3d, 3d' is the letter E, and the black run 3e is the letter F, and the black run 3f is the letter G.
So black run 3g is obtained on one literature, and black run 3h and 3h' are obtained respectively. Between each of the black runs, there are white runs belonging to the own level, and these are arranged as white runs 4-1, 4-2, .
-----4-12. Said black run 3 a to 3 h'
, white run 4-1-4-12, the amount of information (bits) and abbreviations for each are shown in Table 1. Therefore, the total amount of ffI information for # information is the sum of the 9 black run bits and the white run bit. Because it is. There are 140 bits in total. FIG. 1(C) shows the abbreviations shown in Table 1 (black runs are represented by B, white runs are represented by W). The number written after B or W above is
bits), thus visually representing the binary image signal in an easy-to-understand manner.] are arranged according to the scanning progress (from left to right) shown in FIG. 1(b). FIG. 1(d) shows a transmission code that has been subjected to the compression process (one-dimensional run-length encoding process using Wele's encoding code) based on the abbreviations shown in FIG. 1(C). Here, the one-dimensional run-length encoding process means that all bits of the binary image signal are encoded and expressed in a binary scale (hereinafter referred to as run length). It is a bandwidth compression method that reduces the number of
The Wyle encoding code refers to the encoding of the binary image signal using the method shown in FIG. 2 (other encoding methods such as Huffman code and Golomb code have been proposed). , omitted here.) 110110 written at the upper left corner of Fig. 1 (cl) is shown in Fig. 1 (C
), and 010 following the 110110 in FIG. 1(d) corresponds to the W15 in FIG. 1(C).
15 followed by <83. And, the same applies hereafter. Therefore, reference numeral 1 in FIG. 1(d)
10110.010. ----- is the abbreviation W in Figure 1 (C)
i 5 , B 3 , ---1 are respectively encoded. Here, the encoding code of W y 1 e will be explained. In the Wyle encoding code shown in FIG. 2, the encoding code is an address (a code that determines the code length and bias) followed by a reminder (a code that determines the run length). For example, The black run 3a (abbreviation B3) is 3 bits, so to encode it, just look at the right side of "3" in the "white or black run length" column, and if the address is "0". ”, and the reminder is “10”, so it becomes ro l OJ. Similarly, when 10 bits are encoded, it becomes 110001.15
When a bit (for example, the abbreviation W15) is encoded, it is 110
110 (omitted in FIG. 2). In this way, the abbreviation (binary image signal) shown in FIG. 1(C) can be converted into the transmission code shown in FIG. 1(d). (The rWyle encoding code in the present invention refers to the "white or black run length" of 1025 to 1152, M a r
In the gin part. One rOJ is added to each right end of the "address" to form IO bits. ) The amount of information in the transmission code converted by the Wyle encoding code is how many O's or 1's there are. Therefore, the total amount of information in Fig. 1(d) is 91 bits, which is 140 bits in the 1st m(C).
If you compare it with bit, you can see that it is compressed. Generally, the compression ratio is expressed as the total number of pits for one line divided by the total number of bits after compression processing. The compression ratio is 140÷91-1.54. Note that in the transmission code of i1 figure (d), the first block (
110110) shown in the upper left is always regarded as a white run, and since the self-number and black numbers are displayed alternately, the transmission code does not require an identification signal to distinguish between the white run and the black run. , if the transmission code shown in Figure 1(d) is changed to an electrical signal and transmitted, the receiving side will first receive the transmission code shown in Figure 1(d).
Determine how many bits of the own run or black run each code in d) consists of (FIG. 1 (C)), and then create a binary image signal by the above determination (FIG. 1 (B)), Then, based on the binary image signal, a reproduced image [Fig. 1(a)
equal to]. Here, for convenience, converting the abbreviation (z-value image signal) in the tJSt diagram (C) to the transmission code in FIG. 1 (cl) will be referred to as signal conversion A. (This belongs to traditional code conversion). This is because the character shown in FIG. 1 (&) is made up of a combination of line segments with approximately constant width. The bits of the black run change depending on the direction (inclination) of the line segment. Therefore, the black runs are a signal consisting of a relatively short black run (referred to as black run A) to represent a vertical (top and bottom of the figure) or diagonal direction, and a horizontal direction (left and right of the figure).
In order to represent a line segment facing , a signal consisting of a relatively long black run (referred to as black run B) can be divided into 2iliTl. In the threshold values of black run A and black run B, the upper value (M
AX) @9 bits, lower value (MIN) is 3 bits,
Black run A is a black run in the range of less than 9 bits and 3 bits or more, and black run B is a black run other than the above (less than 3 bits or 9 bits or more).
bit or more), then the abbreviation B3 in Figure 1(C)
．． B4 is a black run A, and B9 is a black run B. FIG. 1(e) is the result of removing only the information on black run A from all the information in FIG. 1(c) (consisting of black runs and white runs), and the information on consecutive white runs (for example, W15 and W2) is removed. This means that there existed. In FIG. 1(f), the abbreviations in FIG. 1(6) are encoded according to the WYle encoding code, and furthermore, at the beginning of each code, there is an identification bit (O representing a white run or 1 representing a black run B) ( This is a transmission code with a grained O or l) added. The identification bit was added because it became necessary to distinguish between the white run and the black run B after selectively removing only the black run A. Therefore, the information amount of the transmission code in FIG. 1(f) becomes 74 bits by adding the above-mentioned identification bit.The transmission code in FIG. 1(f) is converted into an electrical signal and transmitted. Receive on the receiving@ side. Depending on the received signal, first, it is determined whether the code block (for example, 0110110) marked with #marked bits represents the run length of white ten or the run length of black run B. . Next, the code is decoded to determine the run length of each run (
From Figure 1 (f) to Figure 1 Ce)], the white run is 2
black run K (
In the black run that is inserted arbitrarily on the receiving side, the black run A (instead of 7, here it is 3 bits) is inserted [from Figure 1 (e) to Figure 1 (g)]. The black runs to be printed have an appropriate run length determined in advance (corresponding to the width of the line segment in the received recorded image, and approximately 0.5
The black run A portion before transmission becomes a black run on the receiving side. To be more specific, Figure 1 (
Black run A in the binary image signals of b) and (c) is. 3ax3cx3dx3d'x3f=3g before transmission
=3h=3h'-83, 3m-3e-84 relationship [first
Figure (C)], but the receiving side made a black run 8' (-B3).
Because it is replaced by . 3ILx3asa3csw3dx3d'm3ex3fx
3g = = 3h bite 3h' Hatchling B3 Figure 1 (g)
It is converted to an abbreviation such as . Therefore, the transmission code of Fig. 1 CF>! If you change it to an air signal and transmit it, on the receiving side,
First, let's determine how many bits of white ten or black run B each code in FIG.
e)), then insert the black run between consecutive 100 runs in the discriminated signal [FIG. 1(g)),
From this, a binary image signal is generated (approximately equivalent to FIG. 1(b)), and a reproduced image (approximately equivalent to FIG. 1(a)) is obtained based on the binary image signal. Here, for convenience, we change the abbreviation (2 (1 image (3)) in Figure 1(c) to the abbreviation in Figure 1(e) and convert it to the transmission code in Figure 1(f). This will be referred to as conversion B. Since the signal conversion B simply aligns the bits of the black run A, the binary image signal reproduced on the receiving side will be changed in the scanning direction (left and right) depending on how the black run N is selected. [Fig. 1(C) is 140 bits, while Fig. 1(g) is 138 bits], this point has been improved, and the scanning direction (left and right) is In order to prevent the bits from changing, instead of simply extracting the black run A from the abbreviations in FIG. You can adjust the difference between black run A and black run A' in advance with the white run after #.
To describe it in terms of A-body, first, black marten 3g in Figure 1(c):
3e are both 4 pits (B4), so in order to make the same bit as the black run 8', one black bit is subtracted to make it 3 pits 1-(B3). The bit is regarded as a white bit, and this is the black run 3a, 3e.
White run 4-2. before (possibly after).
Add to 4-7. (Fig. 1(j) is obtained), and if black run A is removed, the abbreviation shown in Fig. 1(h) is obtained. 1st
In FIG. 1(i), the abbreviations in FIG. 1(h) are encoded according to Wyle's encoding code, and 0 is added at the beginning of each hood.
or a transmission code to which l identification bits are added. Therefore, the information amount of the transmission code in this case is 74 bits. If the transmission code in Figure 1 (+) is changed to an electrical signal and transmitted, the receiving side first determines how many bits of white run or black run B each code in Figure 1 (1) consists of. [FIG. 1(h)] Next, the black run is inserted between consecutive white runs in the discriminated signal [the first
Figure (i)), create a binary image signal from this [Figure 1 (h
)], and a reproduced image [approximately equal to (&) in FIG. 1] is obtained based on the binary image signal. Here, for convenience, we will change the abbreviation (secondary image signal) in FIG. 1(e) to the abbreviation in FIG. 1(h) and convert it to the transmission code in FIG. 1(i). I will call it. According to the signal conversion C, the total number of bits in -scanning becomes <140 bits, which is the same as the abbreviation in FIG. There is nothing to do, the abbreviation (z-value image signal) in Figure 1(C) is 140 bits, Figure 1(f) and Figure 1(i)
Since both transmission codes are 74 bits, the compression ratio is 140÷74-1.89, which shows an improvement in the compression ratio of about 1.54 when compared to the compression ratio (1,54) obtained only by the run-length encoding process. . In this way, the facsimile apparatus of this embodiment using the signal conversion B or the signal conversion C performs the conventional run-length encoding process (the signal conversion A).
It is possible to further increase the compression rate.
Perform IJI. In FIG. 3, the drive section 5 consists of a pulse motor and its drive circuit, and is driven by a document drive signal 19-a, which will be described later.
The original l is fed one line at a time in the scanning line group 2. The transmission document l is nipped and conveyed by a feed roller 6 driven by the drive section 5 and a pressure roller 6' provided opposite to the feed roller 6. An image of characters 7 and the like written on the transmission document 1 is imaged by an optical system 8 on a reading unit 9 (for example, composed of an image sensor such as a COD or a photodiode array and its driving circuit). The reading section 9 starts scanning for one life in response to a reading start signal 18-a, which will be described later. By the scanning, one line worth of image signals is read, and the two lines shown in FIG. 1(b) are read.
When converted into a value image signal, the binary image signal is written to the line buffer 10 via the signal line 9-b in response to a gate signal that has passed through the write clock 9-a, the OR gate 11, and the output 11-a. At this time, 1. ! The read clock 9-a is defeated by 21 by the counter 12. Assuming that the total amount of information on the l line is 2048 bits, the counter 12 is configured to send out a carry signal 12-& when it counts 2048 bits. The carry signal 12-a is
Buffer full number flag flip fist flop (F
/F) is input to the set terminal 13. Further, in the configuration of six or more in which the reading start signal 18-a is inputted to the reset terminal of the F/F 13,
When the read start signal 18-a is sent from the seventh gate 18, the reading section starts reading one line, and in synchronization with the write clock 9-a from the reading section 9, the binary image signal 9-b is output. is sent to the sofa lO to the line reference. The F/F-13 goes into the OFF state when the reading operation of 1 line starts, and turns into the ON state 75 when the reading operation is finished, that is, when all the information of one line of 2048 bits has been written to the line buffer 10. Become. The line buffer lO is the reading speed of the reading section 9 (which is approximately constant regardless of the content of the binary image signal).
and the encoding processing speed of the controller 21 (which varies greatly depending on the content of the binary image signal), and is provided to buffer the speed difference between the encoder and the encoding processing speed of the controller 21, and to smoothly transmit the binary image signal. Although a specific example using one line buffer has been shown here, it is a well-known fact that efficiency can be further improved if two line buffers are used alternately. The controller 21 controls the entire sequence of the facsimile machine, and also performs the signal conversion B or C described in the explanation of FIG. 1 on the binary image signal stored in the line buffer 10. , can be realized not only by microcomputers but also by hardwired control logic circuits. Note that since the controller 21 is the most important part in the facsimile apparatus of the present invention, it will be described in detail later together with the decoder 20. The transmission code created by the controller 21 is transmitted to the First InFirst through the data path of 21-C.
Out (hereinafter referred to as FIFO) is stored in the memory 24. The FIFO memory is provided to buffer the speed difference between the transmission code transmission speed by the controller 21 and the transmission code transmission speed by the modulation/demodulation device (hereinafter referred to as modem) 28, which will be described later. Regarding the characteristics of FIFO memory, for example, Fairchild in the United States
Since it is a product of the company and is described in detail in the MOS LSI3351 catalog and is well known, a detailed explanation will be omitted here. The controller 21 sends the FIRDY sense signal 24- from the 1t6FIFo INPUTREADY terminal (hereinafter referred to as FYRDY) to the FIFO memory 1J24.
a, examine the signal 22-a that has passed through the AND gate 22; At this time, if the FIFO memory is full, FIRDY is set to O.
If there is room to store data in the FF or FIFO memory, FIRDY is turned ON. Therefore, the controller 21
At the time of FIRDY ON. Write the transmission code to the FIFO memory 24 via the data bus 21-C. [At this time, the write clock is set to the FIFO memory +1Jcr+5)I by the shift-in signal 23-a.
It is applied to the IFT N (hereinafter referred to as 1ksI) terminal. ) The transmission code stored in the FIFO memory is sent from the out terminal to the parallel-in serial out shift register 25 via the signal line 24-. The shift register 25 sequentially sends the 8-bit parallel transmission code inputted through the input signal 24-b to the modem 28 through the i-line 25-a.
The transmission clock 28-b, which sequentially takes out transmission codes from the shift register 25 in synchronization with the transmission clock 28-b, is simultaneously input to the octal counter 27 and counted. 8a counter 27
sends out a carry signal 27-a every time it counts eight. The carry signal 27-a is transmitted to the shift register 2.
It is input to the LOAD terminal of No.5. With this configuration, each time the modem 28 finishes transmitting an 8-bit transmission code, a new 8-bit transmission code is stored in the shift register 25.
It is output from the FO memory 24. The carry signal 27-& is also input to the delay circuit 26, and the output 26-a of the delay circuit 26 is sent to the FIFO memory 2.
It is input to the shift out (hereinafter referred to as SO) terminal of No. 4. With this configuration, the shift register 25. After latching the new 8-bit transmission code, the FIFO
A shift out signal 26-a is applied to the memory. The modem 28 consists of a modulation/demodulation circuit and a level adjustment circuit (not shown). Output line 25-a from the shift register 25
The transmission code taken in via the modem 28 is first input to the modulation/demodulation circuit (not shown) in the modem 28 and modulated. A facsimile device transmits an image signal through a telephone line, and the frequency band that can be transmitted by the telephone line is generally from 300 Hz to 3.4 KHz. However, the transmission code from the shift register 25 contains a direct current or OHz signal, and the code cannot be transmitted over a telephone line as it is. Therefore, the carrier frequency is set to a frequency that can be transmitted over a telephone line. However, if the carrier frequency is modulated with the transmission code, the transmission code can be transmitted over a telephone line. (Although the carrier frequency is different, the principle is roughly the same as radio AM modulation and FMyR modulation. Furthermore, PM modulation and other modulation methods can also be adopted.) In addition, when transmitting the above code, Because signals are exchanged (for example, the sending side of the transmission code receives a confirmation signal indicating that the transmission code has been received), the transmission code receives a confirmation signal indicating that it has received the need to demodulate the modulated transmission signal from the other party. (received by the transmitter), it may be necessary to demodulate the modulated transmission signal from the other party. The modulation/demodulation circuit (not shown) performs the above-mentioned modulation and demodulation, converts the code from the shift register 25 into a transmittable signal, and demodulates the 0 code sent from the other party. The output of the modulation/demodulation circuit (not shown) is at the level 1iij (not shown)! input into the node circuit. The level adjustment circuit (not shown) is configured to prevent an adverse effect on the telephone line network caused by an excessive output of a modulated transmission code, an adverse effect on the device caused by an excessive input of a signal sent from the other party, an insufficient output, This is provided to prevent troubles caused by insufficient input. In this way, the modem 28 modulates the transmission code sent from the shift register 25, demodulates the signal sent from the other party, and adjusts the level of the transmission code, etc., to smoothly transmit and receive signals over the telephone line. It shall be carried out accordingly. Then, the encoded signal is sent out from the output line 28-a. Next, the internal configuration diagram of the controller 21 is shown in FIG.
The internal configuration diagram of 080 is shown in FIG. 5. The controller 21 is a microcomputer 8080-
Mainly 21-d, ROM (product name 8316 and below)
, 21-e, RAM (8101) 21-f, 5
TATUS LATCH (8212) 21-1. Bus driver (a212) 21-g, 21-h, 21-
i, and a clock pulse generator 21-k. Also. 14-a, 15-a, and 22-* in FIG. 3 are represented by 21-o in FIG. Regarding the basic operation of the controller and the operation inside the CPU. 1nte1 company's 8080 Microcomputer
SystemsLJ56r'S Manual and other documents provide detailed information, and 8080 software is provided by 1nte1's 8080As sembly.
Language ProgrammingMan
As detailed in the literature of ual et al. A detailed explanation will be omitted here. controller 21;
The interface with the external I10 will be explained below. The address bus 21-b is input to the decoder 20, and various I10 selection signals are generated. The relationship between various I10 and decoder output is shown in Table 2 below. Table 2 As shown in Table 2, each decoder outputs each I
For example, 0 makes a control signal for 10. In FIG. 3, the document drive signal given to the drive unit 5 -
19-a is an AND game of the drive section selection signal 20-c and the WR signal 21-a from the controller 21).
Table 3 shows the I10 control destination created by each gate of the 0th order, which is created as a result of the c logical AND. With the above configuration, the information processing for the 1947 minutes is progressed, and the document drive signal 19 is sent from the controller 21 in FIG.
-a is input to the drive unit 5, the feed roller 6 and the pressure roller 6' feed the transmission document l by one line in the sub-scanning direction (up or down in the figure) [wIJ1 (a) Scanning lines from scanning line 2-1 shown in ? -2] The next line (scanning line 2-2) is read by the reading section 9. Thereafter, scanning is performed in the same manner, and the entire surface of the transmission document 1 is scanned. The flowcharts shown in FIGS. 6(a), (b), and (c) represent the information processing process of the symbol conversions B and C performed in the controller 21, and FIG. 6(d) shows the information processing process at both ends of the scanning line. represents the processing method when there is a black run, and
The path shown in Figure 7 represents the processing route when processing is performed according to the flowchart in Figure 6 (&), and represents that the tree is processed by following the path on the right. . Furthermore, FIG. 8(a) shows the ROM inside the controller.
2Le and FIG. 8(b) show the internal memory map of the RAM 21-f, respectively. First, the symbols used in FIG. 6(L) will be explained below. RL: 16 bits Run length counter 2l-f-i (shown in Figure 8) ML: 16 bits White run length register 2l-f-i (shown in Figure 8)
(shown in Figure 8) BL: 16 bits Black run length register 2l-f-c (
LCNT: 16-bit line counter 2l-f-ho (shown in Fig. 8) D=1 new data memory 2l-d-i (shown in Fig. 5) MOD: Old data memory 2l-d - Mouth (as shown in Figure 5) B: Represents a black bit (=l) Stored in the program in the ROM W: Represents a white bit (=O) MIN=Constant (=3) Stored in the program in the ROM MAX stored in the program: constant (=9) BR: 8-bit operation result memory 2l-f-2 (shown in Figure 8) D: identification bit memory 21-d-c (shown in Figure 5) (shown in Figure 6(a)) Furthermore, the symbol for 口=niko in Figure 6(a) is a sub-
It represents a routine. Regarding ea +t and lololoro]]ko, the sixth
This is explained in Figures (b) and (c), respectively. However, the sixth
The RoroCoroRororo subroutine shown in Figure (C) is the 6th subroutine.
Since it is not described in Figure (a), it will be briefly explained below. The subroutine is as explained in FIG. 2 above.
The FIFO memory 24 has the task of converting the data stored in the RL into the Wyle code shown in FIG. 2 using the Wyle encoding method and transmitting it to the FIFO memory 24 in 8-bit units. A transmission code in the format shown in FIG. 1(i) will be stored. Said Wyle
For a detailed explanation of the encoding method, please refer to Wyle's paper (l(.
SIMLE TRANSMISION BY DI
digital codIg″IRETranS, Com
9, P. 215 (1961-09)), and it is obvious that it can be easily realized by the controller 21 or a device having an equivalent function, so a description thereof will be omitted here. Note that the black runs omitted in (Note 1) in Fig. 6(a) have run lengths (RL) of M
It is a black run that satisfies the condition of IN≦RL<MAX, and (
The calculations performed in Note 2) are integer calculations, and in Note 3, the black at both ends of each line is always transmitted, as shown in the example shown at the bottom of Figure 6(d). Next, in order to explain FIG. 7 and explain the operation of the flowchart in FIG. 6(a), * of the binary image signals in FIG.
The processing will be explained by considering the marked z-value image signal l line. In Figure 6 C&) and Figure 7, first route 29
Start from. When processing one line,
As an initial setting, write O to WL, BL, and BR.
Also, write the negative 1 line bit p to I, CNT, that is, (-16) in the case of 0th order Read B
Execute the it subroutine. [See Figure 6(b)]
, here, the controller 21 outputs the line/buffer selection signal 20-a, and first selects 1 bit of the line < 10 through the AND gate 15 as the binary image signal 15
-a, and write the l bit as new data to new data memory D (21-d-i).
Add +1 to the contents of T ea i
t Finish the work of the subroutine 6 Next, write 1 to RL! 0 Next, go to route 30 and change the contents of D to MOD21-
d- After writing in the mouth, on route 31, ea
it Executes the subroutine. MO after sono
D (21-d-1)),! It is determined whether the contents of =D (21-d-i) are equal. In this case, MOD≠
Since it becomes D, it passes through route 33, and then it is determined whether MOD=W or not. In this case, the answer is YES, so write the contents of RL (in this case RL = 1) to WL via path 37, then write 1 to RL and set LCNT = 0.
Determine whether or not. Since LCNT=-14, the result is No, and the process returns to route 30 via route 50. Therefore, D21-
Write the contents of d-1 to MOD (21-d-port)ea
Execute the d it subroutine. Then M
Determine whether OD=D. This time, the answer is YES, so it passes through route 32, writes 2 to R1, and determines whether LCNT=O. Since I, CNT=-13, the result is No, and the process returns to path 31 through path 35 and repeats the above operation, and
The 6th data counting from the left end of the binary image signal for 1 line consisting of 16 bits is stored in D (2L-d-i), the 5th data is stored in MODC 21-d-port), and R
L=4. When LCNT=-10, the judgment as to whether MOD=D becomes No. Proceed to route 34 through route 33 0th order MOD; Since the judgment of W is NO, proceed to route 36 ゛Here, RL is 3 (MIN) < RL (= 4) ≦ 9 (M
Since the conditions of AX) are met, the answer is YES and route 3
Proceed to 8, here perform integer calculation of (RL-MIN),
Write the difference to BR (21-f-2) 0th order WL =
A judgment is made as to whether or not the current RL (21-
This is done in order to judge whether the black run length whose run length is stored in f-a) is the black run at the left end of one line.Here, it is WL,,l. l of whether LCNT=O or not in the 0th order proceeding to route 40! This judgment is made to determine whether the black run whose run length is stored in RL (21-f-) is the black run at the right end of one line. =-1
Since it is 0, proceed to route 42, write O to BL (21-f-c), and write the result of adding (contents of WL) and (contents of BR/2) to WL. In this case, BR = 1, so As a result of integer calculation, the numbers less than 1 are rounded down and WL=10 tsubo=1. Next, determine whether BR is an odd number or not.Since BR=1, Y
Become ES, proceed to route 45, and take 1 to WL (21-f-exit).
Write +1 = 2, then write O to BR Proceed to path 48 0 Next, write W (=O
) and write WL(21-
After writing the contents of en=], proceed to the loco subroutine [see FIG. 6(C)]. In the encoC subroutine, first it is determined whether RL=O or not.In this case, it is RL-2, so next, the content of DIS (21-d-c) (in this case, information indicating white run) is determined. ’ is stored in DIS) as FI
The process is then sent to the FO memory 24, and then the processing of the above-mentioned mouth [C1 4 RO when 6 or more is 6 or more to execute the loco subroutine] is completed, and the process returns to the main session flow. Next, write EC=1) in the DLS (21-d-c), send the contents to the FIFO memory 24, and then perform the above-mentioned ROROCOLORO] ROSAB. Run the routine. RL (21-f-i) (
Write BR+1), then write O to BR (21-f-2) and BL (21-f-c), and then write LC.
Determine whether NT=α. Since LCNT=-10, N
o, and returns to route 30 via route 51. After repeating route 31 - route 32 - route 35 10 times, R
L=11. With LCNT-0, Mop=o=O, proceed to route 34. Next, it is determined whether MOD=W or not. Since it is YES, proceed through route 37 and proceed to RL (21-f
- Write the contents of b) to WL (21-f-guchi) and RL
It writes 1 to (21-f-i) and judges whether LCNT=O, and since it is YES, it proceeds to route 48 via route 49. Next, repeat the same operation as above, this time, WL=11
．． Since BL=Q, only the white run is RL converted and sent by the e LocoRoco subroutine. Then, proceed to route 52 and MOD=D
Since the determination as to whether or not
Line processing ends. In addition to the processing process for the binary image signal for one line that has been explained above, there are other 23 types of processing process for the 2 (m image signal) shown in Figure @7. The specific processing operations inside the controller 21, such as writing to memory, reading operations, arithmetic processing, size judgment, and match judgment, are described in detail in the two user's manuals of 1nte1. , I will omit the explanation here.Also, if you use a CPU (processing speed, which is superior to 1nte 18080 in terms of processing functions) that is even higher than the CPU explained below for the controller, it will be possible to perform OCR. In the art, it is possible to perform thinning processing, which is a known technique, and if the line width of the image information is made constant by the thinning processing (before performing the signal conversion C),
The applicable range of the signal conversion C can be further expanded. Effects> As described above, the facsimile apparatus of the present invention includes a reading means (reading section a) for reading a transmitted original image, and an encoding system that takes in a binary image signal obtained by the reading means and sequentially compresses and encodes it into a code for transmission. A program control means (controller 21) having a program memory storing a program and a control program for controlling the operation of the reading means, a code memory (FI
FO memory 24), a modulator (modem 28) that modulates the data read from the code memory for transmission. With this configuration, the encoding processing circuit, which conventionally required a complicated hardware circuit, can be simplified, and the encoding method can be changed simply by changing the program. Furthermore, by controlling the reading section using the program control means, reading, which is closely related to encoding, can be performed at accurate timing. 4. Brief explanation of the drawings Figure 1C&) is a front view of the transmitted manuscript mainly consisting of text, and Figure 1(b) is a front view of the transmitted manuscript of Figure 1(a) obtained by scanning the transmitted manuscript.
Figure 1 (C) is the abbreviation for the binary image signal in Figure 2, m1 (d) is the transmission code compressed based on the abbreviation in Figure 1 (C), Figure 1 (e ) is the abbreviation in Fig. 1(c) that has undergone signal conversion B, and Fig. 1(f) is the abbreviation in Fig. 1(e).
), and an identification code is added to the beginning of the transmission code, Figure 1(g) is the same as Figure 1(f)
receiving the transmission code and returning it to the abbreviation shown in FIG. 1(e);
J! An abbreviation with a black run A added between the white runs leading to I, 1st
rI! 1(h) is a symbol obtained by applying signal conversion C to the symbol in FIG. 1(C), and FIG. 1(i) is a symbol obtained by performing compression processing based on the symbol in FIG. The transmission code with ``N'' added to the beginning, Figure 1N), receives the transmission code of Figure 1(i) and returns it to the abbreviation of Figure 1(h). Abbreviations with ゛ added, Figure 2 are Wyle
Figure 3 is a circuit configuration diagram of the facun milli device according to the invention, Figure 4 is a configuration diagram of the controller,
Figure 5 is the internal configuration diagram of the CPU, and Figure 1m6 (a) is
A flowchart diagram of information processing performed by the controller,
Figure 6(b) is the Read Bi in Figure 6(a).
Flow chart diagram of t, Figure 6 (C) is Figure 6 (&)
5enRL flowchart diagram, WIJ6 diagram (
d) is a processing explanatory diagram when both ends of the scanning line are black runs, No. 7
The figure is a processing route diagram of information processing according to the flowcharts in Figures 6 (a) to (C), Figure 8 (&) is an internal memory map diagram of the ROM, and Figure 8 (b) is an internal memory map diagram of the RAM. Figure. In addition, in Figure 1. 1... Transmission document 5... Drive section 6... Feed roller 6... Press roller 8... Optical system 9... Reading section lO... Line buffer 11... OR gate 12... ... Counter 13 ... Buffer full plug flip-flop 1
4.15, 17, 18, 19, 22.23...A
ND gate 16 ... NOT circuit 20 ... Decoder 21 ... Controller 21-d ・CPU (8080) 21-e -
-ROM 21-f... RAM 24... FIFO memory 25... Shift register 26... Delay circuit 27... Octal counter 28... Modem.

Claims (1)

[Claims] (1) A binary image signal representing two levels of brightness obtained by scanning a transmission document, characterized in that one of the binary image signals is partially or completely deleted before being transmitted. facsimile machine.