JPS61176285A - Picture transmitter - Google Patents

Abstract

PURPOSE:To facilitate the detection of a line termination code for expressing the termination of one line in the memory CONSTITUTION:In terms of transmission and reception in a mode G3, the picture data employs the mode of line synchronization, and.

Description

DETAILED DESCRIPTION OF THE INVENTION Technical Field The present invention relates to an image transmitting apparatus that compresses and encodes an image signal and transmits the image signal, and particularly relates to an image transmitting apparatus that has a storage means for storing compressed and encoded image data.

<Prior Art> Conventionally, a facsimile device with an image memory is known as such an image transmitting device.

When storing image data in the image memory, a code indicating the end of one line of image data (hereinafter referred to as an EOL code) is stored together with the image data. This EOL code is also sent together with image data during transmission so that the receiving side can detect the end of one line.

However, the EOL code used during transmission is CCI
An EOL code that complies with the TT recommendations is used, but since this EOL code is somewhat similar to the MH code in terms of bit length, it has been difficult to easily find it in a short time. In view of the above points, it is an object of the present invention to provide an image transmitting device that can easily detect a line end code indicating the end of one line in an image data storage means.

-Z <Example> Below is an example of a facsimile machine that realizes the present invention! An example will be explained in detail.

(mechanism) FIG. 1 shows a sectional view of a facsimile machine.

In the figure, 41 is a COD solid-state line image sensor;
2 is an imaging lens, 43 is a mirror, 44 is a document illumination lamp, 45 is a document feed roller, 46 is a document discharge roller, 4
7 is a document feed tray, and 31 is a document detection sensor for detecting the presence or absence of a document on the document feed tray.

Further, 34 is a roll paper storage cover, 35 is a roll paper, and 36 is a roll paper storage cover.
37 is a paper output tray for originals and recording paper, 38 is a cutter, and 38 is a paper output tray for originals and recording paper.
is a roll paper discharge roller, 39 is a roll paper transport roller, 4
0 is a recording head, and 33 is a roll paper cover sensor that detects whether the cover 34 is opened or closed.

In the figure, when reading a document, the document on the document feed tray 47 is conveyed by rollers 45 and 46.

The document is illuminated by a lamp 44 at the reading position P, and the reflected light is imaged on the image sensor 41 via the mirror 43 and lens 42, and the image sensor 41 converts the image into an electrical signal.

- During actual recording, the roll paper 35 is connected to the roller 39 and the head 40.
An image is formed on the heat-sensitive roll paper 35 by the head 40 at the same time as the heat-sensitive roll paper 35 is conveyed. When the recording of -page is completed, the roll paper 35 is cut by the cutter 37.

The paper is discharged onto a paper discharge tray 36 by rollers 38 .

(Basic Block) FIG. 2A is a basic control block diagram of the facsimile apparatus of this embodiment. In the figure, 1 is a reading unit that reads an original image and converts it into an electrical image signal; 3, 5, 7; One aspect of this is a random access memory (hereinafter referred to as RAM) that functions as a buffer for temporarily storing the image signal;
9 is a first-in-first-out RAM (hereinafter P
IFORAM), 11 is a read-only memory (hereinafter referred to as ROM) that stores the operating program of the MPU 23, and 13 is an RA that stores flags, data, etc. necessary for the operation of the MPU.
M, 15 is an operation unit having an input key 1 display, etc.; 17 is a recording unit that records a copy image, a received image, and management data on thermal paper; 19 is a modem that modulates transmitted data and demodulates received data; 20 2 is a telephone, 21 is a network control unit (hereinafter referred to as NCU) for controlling the connection of the communication line 22 to the modem 19 or telephone 20, and 25 is for transmitting, in addition to the original image, a call stamp and the name of the sender as image data. 23 is an MPU that controls the entire system. In this embodiment, the MPU 23 uses an Intel 8086 that can directly access a 16-bit data bus 24 and a memory space of up to 4 megabytes.

The advantage of using this MPU is that the 16-bit
data bus, it became easy to handle encoded image data. For example, to handle 2048 bits of data with a run length code, 12 bits are required.
data is required, and if you use an 8-bit MPU, you will have to access it twice, but if you use 16-bit data, you will have to access it twice.
You only need to access it once.

Furthermore, since a large capacity memory space can be directly accessed, it has become possible to use the system memory as an image memory and provide a reporting function. Conventional devices use an external memory unit or a memory that cannot be directly accessed by the MPU via the bus even if it is inside the device to provide a relay function as image memory, but this increases the complexity of the circuit and the device. There were problems such as increasing the size.

(Functions of MPU) The basic functions of MPU 23 include 6 as shown in Figure 2 CB).
There are seeds. Each function will be explained below.

Encoding function - (Run length → MH, MR code conversion, etc.)a
) When performing run length → MH coat conversion encoding, first, a one line read command is issued to the reading unit 1. Then, the reading unit 1 converts the read image data for one line into a run length code and writes it into RAM3, and the MPU 23 reads the run length code from RAM3 and uses it to convert the code in ROMI 1. Look up the table and convert it to MH code. The conversion table is developed on ROMI1, and the run length code is used as an address, and MH code data corresponding to the run indicated by the address is written. The structure of the MH code data is shown in FIG.

In Figure 3 (A), MH is placed left-justified in the upper 12 bits.
Enter the code. Furthermore, since the MH code is a variable length code, the code length information of the MH code is entered into the lower 4 bits to recognize the code length. M in the upper 12 bits
Although the H code is assigned, the maximum 1 in the MH Coco table is
A 3-bit code exists. To deal with this, if we focus on codes with a code length of 13 bits, we can see that the beginning (MSB) of all codes starts with "0". Therefore, the data in the conversion table is
The 12 bits excluding II are used as the MH code, and information with a data length of "'13" is added. Then, when the conversion table is checked and the data tone is "13'", M
A method is used in which the PU adds "0" to the beginning of the code.

In this way, all MH codes and their code lengths fit within 16 bits, so a 16-bit MPU (
The MH code can be searched for at high speed.

b) Run ψ length → MR coat conversion Conversion to MR code is performed by the MPU 23 with reference to the basic flow shown in CCITT's T4 recommendation, but the amount in the basic flow is also a frequent and important item. One example of this is "detection of pixel white/black reversal." Therefore, the data that the reading section writes into the RAM 3 is converted into a run-length code so that the detection can be easily performed.

A program flow for converting a run-length code into an MR code is shown in FIG. 3(B), and a subroutine for determining the parameter bl is shown in FIG. 3(C).

In Figure 3 CB), first set the parameter a O+b1 to O
By initializing to
After determining b2 in the routine of
The decision is made by calling the code. MR of T4 recommendation
At the same time as the MR code is determined in the encoding routine, the next value of the parameter aQ is determined.

In FIG. 3(C), in the target line where parameter b1 is on the right side of aO, aO is the color (white).

It is determined according to the definition of the Recommendation that the first color change point is different (black).

Since this conversion to MR code is performed from the run-length fist code, it can be performed much faster and more easily than when converting from raw image data.

C) CG code → MH code conversion This device has the function of converting information such as characters in addition to the image data read by the reading unit into MH code and transmitting it as image data. , first CG
With the code, extract the raw data corresponding to the CG code from CG25, convert it to a run length code,
Furthermore, it is converted into an MH code and transmitted. The reason why the conversion table output is raw data instead of run length code is because if you create a table using run length, the number of codes will increase and a large CG child table will be required. This is to reduce capacity, and by using raw data, there is also the advantage that decoding is not required when transmission is in uncompressed mode such as G2 mode.

d) Handling of EOL When transmitting and receiving image data in G3 mode, line synchronization is used for image data, and EOL is used as a line synchronization signal for this purpose.
OL (End OF Line) is set. EOL is 11 consecutive “□11 plus 1 (MR
In this case, l or O is added. ).

Each time the MPU 23 detects the end of one line, it adds this EOL to the image data and sends it out. When adding this EOL, it calculates the transmission time of the transmission line and determines if it is less than the minimum transmission time. In such cases, EOL is added after inserting fill pits to achieve the minimum -transmission time. In actual transmission, the MH code is a temporary FIFO.
The code is stored in the RAM 9, and the MPU reads the code from the RAM 9 and transmits it. The calculation of the minimum transmission time and the insertion of fill bits are performed when the MPU 23 reads the code from the RAM 9 and transmits it. Therefore, detection of the line end signal EOL of the poem to be read from the RAM 9 becomes an important issue. Therefore, in this device, the EOL of the poem read from RAM9 is
The following method is used to simplify detection and EOL transmission.

First, the basic philosophy of handling EOL is (1) E OL is added when writing to RAM9.

(2) EOL detection when reading from RAM 9 is 2
Perform bytes continuously.

(3) When sending data from the RAM 9, the second byte of O out of two consecutive O bytes is not sent.

There are three points. I will explain the following two cases separately.

The storage format of data and EOL in RAMQ when l'' in the final data of a line exists in the final byte is explained in the fourth section.
As shown in the figure. In the figure, the final byte A's I)T is image data. Put an O after the data DT in Pi)A, set O to all B and C, and insert IX at the D-th byte. However, depending on the number of O's inserted in byte A, the number of zero's before 1x of the D-th byte is changed as shown in the table below.

In this way, even if the EOL 0 in the memory is sent out by 1/<ite, 11 O's can be secured.

Next, FIG. 5 shows the storage form of data and EOL in RAMQ when 1°' of the final line does not exist in the final line. As shown in the figure, when the data DT included in the final pie A is all 0, all O's are inserted into the remainder of byte A, and O's are also inserted into the next byte B. And part-time job C
After inserting the number n of O's inserted into the byte from 1, insert 1x.

Since this does not exist, the number of 0s inserted into the A-th byte does not take into account 4 or less.

Furthermore, in consideration of white line skip transmission, as a margin judgment criterion, if one line is all white data, O in the second byte is distinguished as O1' (hex display).

By writing to PIFORAM 9 in the above format, EOL detection when reading from RAM 9 will be 2 consecutive 0 bytes or 1 byte 0 and ``0'' (hex).
It can be easily done. Furthermore, when transmitting the read data, EOL can be easily transmitted by deleting the second byte O (or "o i ").

Although it is possible to send EOL without deleting the 0 in the second byte, deleting it prevents unnecessary data from being sent and reduces the transmission time.

Decoding function (MH, MR code → run length code) a) MH code → run length code conversion Encoding method is FI
Take out the MH code from FORAM9, MH→
The decoder is performed using a run-length conversion table, but the method of drawing the table is somewhat different from the run-length to MH code conversion method described above.

FIG. 6 shows a conversion flow from an MH code to a run-length code, and FIG. 7 shows a table. As is clear from the flowchart in Figure 6, the MH code is searched bit by bit, and if it is 0, the data at the address indicated by the current address pointer is searched, and if it is 1, the data at the next address is searched. If it is ``1pa'', the data is the run length, and if it is ``o'', the data is written to the address revo pointer and used for the next search.In other words, the MSB
The MH code is searched bit by bit until it finds "1'° data (data starting with 8).Figure 7 MH code black" OOOO111
) Shows a search example. It can be seen from the figure that the aforementioned code is a run length code of "black 12".

And the conversion table is made up of different black and white codes. The reason for this is that the same MH code exists with different run lengths for black and white.

b) MR code → run length conversion A table search method similar to MH → run length conversion is performed using a conversion table, but the data with MSB = 1 is written not as a run length code but as the program jump address. ing. And on that flight,
Processing corresponding to the MR code is performed to generate a run length code.

MR encoding is an encoding method using two-dimensional compression, so 1
There is no run length code corresponding to one MR code. Since the run-length code must be created using the MR code based on the data of the previous line, there are no entries in the table. Figure 8 R-MR code ゛OOOOl
1” is shown below.

(Calculating the minimum transmission time and inserting and deleting tFu) EOL is added after one line of data to the G3 transmission poem, but the transmission time of the 1 life's worth of data sent at this time is calculated. , if it is less than the minimum transmission time, Fll
1 bit (data 0) is inserted and EOL is added after the minimum transmission time is exceeded.

This device determines whether the transmitted data exceeds the minimum transmission time by converting the transmitted data into the number of bytes based on the minimum transmission time and transmission rate, and then determines whether the number of transmitted bytes is equal to or greater than this converted number of bytes. There is.

The number of bytes transmitted within the minimum transmission time is 9800x10, assuming the minimum transmission time is LOMS and the transmission rate is 9600BPs.
X10-3 = 12 [bytes].

The Fi1M bit is inserted in byte units.

In this device, transmission/reception data in the G3 mode and data stored in the memory are always transferred via the FIFORAM 9. If Fillit, which is an idle signal, is stored in the RAM 9 as image data, the RAM 9 will be wasted.

Also, since the apparent number of Fill bits changes depending on the capability of the other device when performing memory transmission, when storing memory, the maximum number of Fill bits calculated from the maximum possible minimum transmission time and data speed must be inserted. It will stop happening.

Therefore, in this embodiment, during G3 mode transmission and memory storage, no Fill bit is inserted into the PIFORAM 9, and after reading from the FIFORAM 9 during transmission,
FiJllQit is inserted and sent.

Furthermore, during reception, a method is used in which if three or more bytes of 0 occur consecutively, the 0 bytes after the third byte are not written to the RAM 9.

(Fine→Standard Conversion) This embodiment has a function of performing fine→standard conversion when transmitting image data stored in the FIFORAM 9 in the MH code. Comparing fine and standard, the line density in the main scanning direction is equal to ape sentence/mm,
The line density in the sub-scanning direction is fine 7.7 ine/mm,
The standard is 3.85 Jljine/mm, which is 1/2 of the fine. Since the data stored in the PI FORRAM 9 is separated into one line by EOL, the line can be easily determined. Therefore, in this device, when transmitting data from the PI FORRAM 9, fine to standard (scanning line density) conversion is performed by transmitting every other line.

FIG. 8(B) shows a processing flowchart when a data request interrupt is received from the modem 19 with and without scanning line density conversion.

First, when Interabdo enters, from FI FORRAM9,
Recalls the data of the current read address pointer. If the data is not EOL, after outputting the data to the modem, the pointer is moved to 11, and data transfer is repeated.

When an EOL is detected, as described above, convert the EOL in RAMQ to an EOI for transmission (CCITT recommendation), then add a fill bit if necessary, and send the EOL and fill to the modem. Output to. Then, it is determined whether fine→standard conversion is necessary, and if not, the pointer is increased by 1 and reading of one line of data is completed. −
If it is necessary to convert the force scanning line density, the address pointer is advanced to the next EOL, data for -lines is deleted, and then the process returns to the main routine.

(Run length → raw data conversion) When transmitting memory in G2 mode, FIFORAM9
The data stored in the MH code must be transmitted as raw data. In this device, the data conversion is performed by software, but it is quite difficult to convert directly from MH code to raw data. Therefore, the program is simplified by using the decoding function described above to convert the MH code once into a line-length code, and then converting it into raw data.

For example, the conversion from run length code to raw data is
This is done as shown in Figure (C).

That is, the RLC code is read, and if RLC is a black date, 1° is outputted to the line memory, and the process is repeated until 1RLC becomes O. If RLC is white data, 0° is output to the line memory and the same process is repeated until RLC becomes 0, thereby converting RL+RAW.

(B4 → A4 reduction by software) In this example, 2
Reading is performed using a reading section 1 having a 0.048-bit light receiving element.

Therefore, it is possible to transmit a B4-width original at 8 pel/mr. However, if the other machine only has the recording capacity of A4 width) data of B'4 (2048
There is a need to convert the A4 size data (1728 bits) into A4 data (1728 bits) and transmit it. In the case of normal document transmission, the processing is carried out in the reading unit 1 using optical or electrical means, but when considering memory transmission, the reading unit 1 has a reduction function that is also considered from the data flow. It is impossible to use. Therefore, in this embodiment, reduction is performed using software.

First, the data stored in RAM9 as MH code is converted to run-length code using the decoding function, and then the data is reduced in the main scanning direction of the line, and converted back to MH code (or raw data in the case of G2). Convert and transfer to modem.

Note that the reduction in the sub-scanning direction is performed by thinning out data line by line, as described above.

do.

The number of dots in the lifetime scanning line of B4 is 2048 dots, A
4 is 1728 dots. If you factor this out, it becomes 32
X26: 27X26 gives a ratio of 32=27. So, the data of 2048 dots of B4 is 6 by 32 dots.
Divide into 4 blocks. Then, for the 32 dots in the l block, it is sufficient to thin out 5 dots and convert it to 27 dots. 1 block 3 in Figure 8 (D)
2 dots are shown. 6, 13, 19 with diagonal lines in this diagram
, 26. If the 32nd dots are thinned out, the dots can be thinned out at a substantially uniform density in the main scanning direction.

FIG. 8(E) shows a flowchart for performing this conversion. In order to facilitate the explanation of the flowchart, for example,
The data as shown in figure (F), ie the run length code, is white 8.
Black 5. An example of converting a 32-dot code of white 15 degrees and black 4 to 27 dots will be explained.

First, the total RL counter for one line is TCNT in SPI.
, 32-dot i counter TRL, post-conversion run length code SRL are set to O, the decimation number counter MC in 32 dots is set to 5, and MA indicating the address to be decimated is set to 6.
Set to .

Then, in SF3, the first RL code white 8 is called from RAM9, and both TCNT and TRL are set to 8 in SF3. Since TRL=8 is larger than MA=6, White 8's data RLC is converted to White 7's data RLC (S
P5).

RLC = white 7 and SRL is O, so MA is 13 with 5 PIO
Then, MC becomes 4 and returns to SF3 again. This time TRL
=8 is smaller than MA=13, so proceed to 5P16 and SR
Since L is set to white 7 and TCNT is less than 2048, the next RLC = black 5 is called and T
Both CNT and TRL are 13. TRL is MA=1
Since it is equal to 3, SF3 and RLC become black 4. and S
At F3, the colors of 5RL = white 7 and RLC = black 4 are different, so S
At F3, white 7 data is output to the line memory and SRL is reset to 0. Furthermore, MA is set to 19 and MC is set to 3, and the process returns to SF3 and proceeds to SPI 6. This time, RLC=black 4 is set in SRL. Then the next RLC = white 15 is called, TCNT, TRL
is set to 28.

Since 28 is greater than MA=19, RLC=white 15 is converted to white 14, and in SF3 5RL=black 4 and RLC=white 1
White of 4 is compared, data of black 4 is output to the line memory, and SRL is reset to 0.

Then, MA is set to 26 and MC is set to 2. At step SP4, since TRL=28 is still larger than MA=26, the data of White 14 is further converted to White 13, and at this time, S
Since RL is 0, MA is set to 32 and MC is set to 1 at 5PI0.11 without determining and outputting SF3 and SF3.

Return to SF3 again, this time MA=32 has TRL=2
Since it is larger than 8, white 13 is set in SRL at 5P16. Then, after calling the next RLC = black 4, SF3
Then, white 13 is output, and the edge point 3 is output in the same way.

As mentioned above, in the upper row of FIG. 8(F), white 8° and black 5. white 1
5. The data for Black 4 is at the bottom: White 7° Black 4. White 13. It is almost evenly converted to the black 3 run length code.

Note that step S13. S14, 5p15 is l block 3
This shows the initialization of MC, MA, and TRL when 2-dot processing is completed. In particular, 5P15 also has an adjustment function when the run length code spans between blocks. or,
5P18 indicates the output of the last run length code of one line to the line memory.

In this way, it is possible to convert the number of main scanning dots while using the run length code.

(Operating Modes) As shown in the table below, there are a large number of operating modes regarding the transmission, reception, and transfer of image data in this embodiment. The data flow and code format in the following mode will be explained using figures.

First, this device operates in the 14 operation modes M1 to M14 described above.
Flowcharts of the determination algorithm of the MPU 23 used in determining are shown in FIGS. 9(a) to 9(C).

In this embodiment, activation is performed using the start key 51, one-touch dial key 54, speed dial key 53, and memory key 52 on the operation panel 50 shown in FIG.

Furthermore, the judgment is made based on the outputs of the sensor 31 that detects the presence or absence of a document in FIG. 1, the sensor 32 that detects the ON10FF state of the telephone hook, and the roll paper cover sensor 33.
A branch is made.

Furthermore, the mode of the other party is set to G due to the communication of pre-procedure signals prior to facsimile message (image data) communication.
You can know whether it is 3 mode or G2 mode. At the same time, it is also possible to know whether the other party's device has an MR encoding function or only an MH encoding function.

Furthermore, it can be determined whether the PIFORAM 9 can be used during message communication based on the usage state of the image memory of the device itself. If memory is stored in RAM9, RAM
9 cannot be used, and if memory is not stored,
RAM9 can be used.

The 14 operation modes determined by this flow are marked with M1 to M14. First, when the start key is pressed, the handset goes off-hook as shown in Figure 9(a). If it is on-hook, if the document is in the transmission position, the mode shifts to document copy mode M14, and if there is no document and the roll paper cover is closed, the roll paper cutter operates, When the cover is open, a predetermined amount of roll paper is dispensed.

- On the other hand, in the case of off-hook, if there is a document, the mode is set to transmission, and the mode shifts to M1, N2, N3, and N6 depending on the mode of the other party's machine and whether the RAM 9 can be used. If there is no document due to off-hook, the reception mode assignment shown in FIG. 9(b) shifts to the routine. In FIG. 9(b), M7 to Mll are selected depending on the partner machine mode and the availability of RAM9.

FIG. 9(C) shows a routine for mode distribution when the memory key 52 is pressed.

When the memory key 52 is pressed, a software timer is activated, and when a document is placed on the reading section l during this timer, the mode shifts to the memory storage mode M12, and the image data of the document is stored in the RAM 9.

When the start key 51 is pressed when no document is placed in the reading section l, the mode shifts to the memory copy mode MI3 in which the image data in the RAMQ is recorded in the recording section 17 if it is an on-footer.

Also, if it is off-hook at this time, it shifts to memory transmission mode. When the one-touch key 54 or speed dial key 53 is pressed, the mode shifts to memory transmission mode regardless of the hook state. Memory transmission mode is 02 or 0 when the other machine is
Depending on whether there are 3 machines, G3 memory transmission mode 1M4. Or it is assigned to G2 memory transmission mode 1M5.

Also, when the memory key is pressed and no original is placed in the reading section and no other key operations are performed, the amount of image data accumulated in the RAM 9 is displayed on the display 55 (Fig. 10), and the software timer is activated. waits for the timeout and returns to standby mode.

The flow of image data according to each mode M1 to M14 will be explained below.

(Mode Ml) The flow of image data in G3 original transmission, MH, and RAM9 available mode Ml will be described with reference to FIG.

The reading unit 1 reads 1247 in response to a reading command from the MPU 23.
Convert the image data of minutes to run length code RL and R
Write to AM3. Then, the MPU 23 transfers the data in RAM 3 to two tree line buffers RAM 5 and RAM 7.
The run length code RL read from the two line buffers is encoded into an MH code and written into the PIFORAM 9. And M
The PU 23 transfers the MH code one byte at a time from the FIFORAM 9 to the modem in response to the inter-data request from the modem 19. Also, at this time, the minimum transfer time is calculated for each line and fill bits are inserted.

In addition, character information such as the source and time of transmission to be added to the beginning of the image is converted to PI FO using the raw image data 25 output from the CG 25 using the raw data → MH code conversion function.
Transferring to RAM9.

Reading unit 1 + RAM 3 and modem 19 → NCU 21 in the diagram
All other data transfers are performed via the bus 24 of the MPU 23 except in the case of .

The inter-interval of data requests from the modem 19 changes depending on the transmission rate.

Data transfer is done in bytes, so 9600
For bps, 8/9600=0.83XIO-3se
An interwoven occurs every c.

Furthermore, when the data transfer from RAM3 to RAM5 and RAM7 is completed, the MPU 23 outputs a reading command to the reading section. While the MPU 23 is performing the encoding process ENC1 and the interwoven process, the reading unit 1 performs reading of the original and converting raw data to run length data.

(Mode M2) G3 original transmission, MR, RAM9 available FIG. 12(A) shows the flow of image data.

The data flow is almost the same as in mode M1. The difference is that the code after ENC23-1 is an MR code. However, the data from CG25 is output from ENC23-1 in MH code. For example 24X
When adding a 16-dot character to the beginning, 24 lines of data are transmitted using the MH code.

FIG. 12 (B) shows a program for storing CG data in the MH and image data in the RAM 9 in the MR. First, the number of lines of CG data is initialized, the data of each line is called from the beginning, the raw data is converted to a run length code, the RL code is converted to an MH code, and each line is stored in the RAM 9.

When the 24th line is completed, the image data of the RL code is read from RAM 5 or 7, as shown in FIG. 3(B).
, (C), each line is MR encoded according to the MR encoding routine of (C).
It is converted into a code and stored in the RAM 9.

(Mode M3) FIG. 13 shows the flow of G3 original transmission, MH, and RAM9 usable image data. RAM in Figure 11
Unlike the case where MH5 code can be used, RAM7, which was used as a line buffer, is used as a buffer memory for the MH5 code.

Therefore, there is only one line buffer in RAM5.1, and encoder ENC23-1 can only handle data for -line, so MR transmission cannot be performed if RAM9 is unavailable.

The reason for this is that MR encoding requires line buffers for two lines: the current encoded line and the reference line.

(Mode M4) 03 Memory Transmission M H-----Fig. 14 (A), (b), (c) The flow of image data in the case of mode M4 is shown in Fig. 14 (A). The PIFORAM 9 stores image data read in fine mode or standard mode in the form of MH code. Further, various information of the image data is stored as a label at the top of the page, as shown in FIG. The information includes the read size (number of main scanning dots) SZ of the image data, whether it is fine or standard (scanning line density) F/S, the number of EOLs PFN of the page, etc.

Therefore, if the size of the recording paper of the other machine is smaller than the reading size SZ, it is necessary to perform the main scanning dot number conversion described above, and even if the recording paper size of the other machine is stored in RAM 9 in fine mode, If only the standard mode is available, it is necessary to perform the scanning line density conversion described above.

FIG. 14(B) shows the distribution routine. In FIG. 14(B), first the EOL counter EO
Set C to O, and set the recording paper size A of the other machine in the previous step.
Sense SZ. And compared with label SZ, ASZ
is greater than or equal to SZ, select mode M4-1 or M4-2. In this case, there is no need to convert the number of main scanning dots.

Also, if ASZ is smaller than SZ, mode M4-3
．． M4-4 is selected. In this case, it is necessary to convert the number of main scanning dots.

And, the other machine does not have a fine recording mode, and the RAM
9, it is necessary to further convert the sub-scanning line density, and mode M4-2) or M4
-4 is selected.

That is, M4-1 does not require either main scanning dot number conversion or sub-scanning density conversion. M4-2 requires only sub-scanning density conversion, and M4-3 requires only main-scanning dot number conversion. Also, M4-4 is necessary for both conversions.

A detailed explanation of the data flow in each mode will be given later, but when the transmission of one line is completed, mode M4-1.4
-3, EOI, counter EOC is +1, M4-2.
In 4-4, increase EOC by +2. And EOC is RAM9
The process moves to a page end subroutine that matches the PFN indicating the EOL number of that page.

The page end subroutine is shown in FIG.
The label GE indicating the final page of the group stored together with a series of pages in the MQ is checked, and if the page is the final page of the group, an EOP indicating the end of transmission is output to the other party's machine, and the transmission ends.

On the other hand, if it is not the last page of the group, the next page SZ, F/
S is read out, and if F/S and SZ are the same as the previous page, an MPS signal is output indicating that the next page will also be sent in the same mode. If not, E indicates to repeat the previous step from the beginning.
It sends an OM signal to the other party's machine.

The distortion of image data in each mode of M4-1 to M4-4 will be explained below.

(M4-1) Main scanning dot, no sub-scanning line density conversion The image data in the RAMQ is added with a fill bit by the Fi statement 23-3, and sent from the NCU 21 via the modem 19. Also, the output raw data of CG25 is ENC23-
1, it is MH encoded and is not directly transferred to the FiJ1 statement.

(M4-2) The MPU 23 with sub-scanning line density conversion converts the MH output of RAM 9 from fine to standard using F/523-4 while keeping the MH code, that is, deletes every other line of data, and converts the MH output from RAM 9 to standard using the F/523-4, and deletes data every other line. Output to 7. RAM3.5.

The MH data in MH 7 is added with a fill bit by Fiuu 23-3 and transferred to modem 19.

Also, the output raw data of CG25 is also ENC23-1 and RA.
It is output to the Fi report 23-3 via M3.5.7.

(M4-3) MPU23 with main scanning dot number conversion is R
Extract the image data of Ml (from AM9, DEC23
-2, it is converted to run length code RL, and in the state of RL, conversion from B4 to A4 is performed. RA that is used as a FiFoi memory and returned to MH code with ENC23-1
Output to M3.5.7. Thereafter, a fill bit is added to the signal by Filn 23-3, and the signal is transferred to the modem 19. CG
The output raw data of No. 25 is also converted into MH code by the ENC 23-1 and then transferred to the Fi Bunmemi 233 via the RAMs 3 and 5.7.

(M4-4) The MPU 23 with both conversion converts the MH data in the FIFORAM 9 into M
After F/S conversion as H, and further converting to DEC23-2 run length code RL, B 4/A 4 conversion, and converted run length code RL to E and NC23-1 to M.
Return to H code and transfer to RAM3.5.7. CG2
Similarly, the output of 5 is ENC23-1, RAM3.5.

7 to the FEII.

(Mode M5) G2 memory transmission-111 Figure 15 M
The PU 23 extracts the MH code from the FIFORAM 9, decodes it into a run-length code RL, converts it to raw data RAW, and transfers it to the RAM 5.7 alternately one line at a time. Then, raw data is sequentially extracted from the RAM 5.7 and transferred to the modem 19. In addition, when performing mode conversion from fine to standard, RAM9 and DEC23-
When performing reduction, B4/A423-5 conversion is performed between two DEC23-2s.

The output data of CG25 is stored in RAM5 in the form of raw data RAW.
．． 7 to the modem 19. However, in this case, the CG25 data is transmitted at 7.7 JLin, e/mm in the sub-scanning direction without thinning out the scanning lines, so that the character size is doubled vertically compared to the G3 mode. This is done to ensure that the sender information can be read even in G2 mode, since G2 is an analog transmission and the image quality deteriorates significantly due to transmission.

(Mode M6) G2 original transmission --- Figure 16 All data is transferred in the form of raw data. Reading part l
writes one line of image data as raw data to the RAM 3 in response to a read command from the MPU 23. And MPU2
3, data in RAM 3 is alternately transferred one line at a time to two line buffers RAM 5 and RAM 7 as is. Then, in response to data requests interrelated from the modem, raw data is transferred one byte at a time from RAM 5 or RAM 7 to modem 1.
Transfer to 9.

Further, character information such as a source record added to the beginning of an image is transferred from the CG 25 to the RAM 5.7 as raw data.

Furthermore, in the case of G2 mode, 1728-bit image data including a synchronization signal is written into RAM5 and RAM7. The MPU 23 creates an image signal corresponding to this synchronization signal.

(Modes M7, M8) G3 reception MR mode, RAM9 usable (unavailable) - 111 Figure 17 When the MPU 23 receives the MR code from the line via the NCU 21 and modem 19, it first deletes the fill bit, RAM9
If there is no data in RAM9, it is transferred to RAM9, and if there is data in RAM9, it is transferred to RAM3 as is. and R
The MR code is sequentially extracted from AM9 or AM3, decoded into line length code RL, and then transferred to RAM5 and RAM7 alternately one line at a time. At the same time, the run length code RL is transferred to the recording section 17 and recorded.

The decoded run length code RL is stored in RAM5, R
The reason why it is transferred to AM7 and stored is to use it as front line information when converting into MR code.

(Mode M9.MIO) G3 received MH code RAM 9 usable (unavailable) - 111 Figure 18 When the MPU 23 receives the MH code from the line via the NCU 21 and modem 19, it first deletes the fill bit and then stores it in the RAM 9.
If available, go to RAM9, otherwise go to RAM3.5.
Transfer to 7 as MH code. and RAM9 or 3
, 5.7, the MH codes are sequentially extracted, converted into line length codes RL, and transferred to the recording section 17 to be recorded.

(Modem Mll) G2 reception; ---1 Figure 19 In G2 mode, uncompressed raw data is sent, so
3 receives the raw data from the line via the NCU 21 and the modem 19, and then alternately stores the data one line at a time in the line buffer RAM.
5゜Transfer to RAM7. And RAM5, RAM
7, the raw data is sequentially extracted and transferred to the recording section 17,
Record.

Further, a 1728-bit image signal for one line demodulated by the modem 19 is written into the RAM5 and RAM7.

Since this includes the image signal obtained by demodulating the synchronization signal, the MPU 23 removes the image signal corresponding to the synchronization signal when transferring it to the recording section 17.

(Mode M12) Memory storage - 111 Figure 20 FIF
It is almost the same as mode M1 until it is stored in ORAM9 with the MH code, and the difference is that there is no data from CG25, and that a label LB for file management is added to the top of the page from RAM13 when transferring to RAM9. It is.

Let me explain about labels here.

The label is composed of 24 bytes as shown in FIG. The 1st to 3rd bytes contain LPM indicating that the labeled data is the final page and NPA indicating where the start address of the next page is. The fourth byte contains information for each page. In the fourth byte MSH, when data is divided not only in page units but also in group units, information GE indicating whether or not this is the last page of the group is entered.

For F/S, the scanning line density is standard (0.85 lines/mm
) or fine (7 lines/mm).

The data in RAMQ is MW and MR in MD.

Information on whether the data is stored in RL, RAW, or ASCI I code is entered. SZ contains information as to whether the data in RAMQ has a reading width of A4, B4, or A3.

The fifth byte is GPC, which indicates the page number within the group when the data is divided into green leaves. The 6th to 9th bytes contain the total number of lines PLN for the page, the 10th to 14th bytes contain the time when memory storage was performed, the 10th byte contains the minutes, and the 11th byte contains 1 o'clock. The 12th byte is “
The 13th byte stores "January", and the 14th byte stores "Year". Furthermore, the file name PFN of the page is entered as a code in the 15th to 24th bytes.

Then, during memory transmission and memory copying, the mode is determined and information is added based on the information in this label. Regarding time data, when memory copying is performed, the time at the time of memory storage is used as a header based on the information in the label. The information in the label LB is ignored and the transmission time is sent when the label is printed and sent to the memory. When time-specified transmission is performed, consideration is given so that the time printed on the received image is not the time stored in the RAM 9, but the time at which the transmission was actually performed.

Furthermore, the image data and label LB once stored in the RAM 9 are cleared manually or automatically by the operator. The automatic clearing flow is shown in FIG. 22.

Note that memory clearing is not performed after memory copying.

(Mode M13) Memory copy-111 Figure 23 (A)
The MPU 23 sequentially extracts the MH codes from the RAM 9, converts them into run-length codes, transfers them to the recording section 17, and records them. Further, the header information is converted from character code to raw data via the MPU, transferred to the recording unit 17, and recorded.

The time in the header is the time when memory storage was performed in the file management label LB stored in RAM9.
25 into an image and recorded in the recording section 17.

FIG. 23(B) shows the time management subroutine. First, in the case of transmission mode, the clock 27 (
The date and time data shown in FIG. 1) are output to the CG 25, and the transmission time is transmitted together with the image. At the same time, RA for communication management
The time and the destination number are stored in M13.

Also, during memory copying, the date and time data TD in the label
is output to CG2'5. At the time of memory storage, the date and time data of the clock is output to the RAM 9 as data TD. Also, when receiving data, the data of the clock 27 is transferred to the RA″M.
13 along with the location number of the other party. Note that in the case of original copy mode, □ time data is not involved at all.

(Mode M14) Original Copy --- FIG. 24 The reading unit 1 receives a reading command from the MPU 23 and writes data for 1 life into the RAM 3 in the form of raw data RAW. Then, the MPU 23 sequentially extracts the raw data from the RAM 3,
The data is transferred to the recording unit 17 and recorded. The output data of the CG 25 is transferred to the recording section 17 in the form of raw data and is recorded.

<Effects> As explained above, the image transmitting device of the present invention includes a reading means for reading an image and converting it into an image signal, an encoding means for compressing and encoding the image signal, a storage means for storing encoded image data, In an image transmitting device comprising transmitting means for transmitting image data in the storage means, the storage means stores a line end code indicating the end of one line of the image data in a predetermined format, and the transmitting means The end code is sent together with the image data in a format different from the predetermined format.

Such a configuration facilitates processing associated with line termination of the transmitting means. In particular, by setting the code length of the line end code in the storage means to be longer than the maximum code length of the image data, or by setting it to be longer than the number of parallel processing bits in the transmitting means, the line end code can be set very easily. can be detected.

[Brief explanation of the drawing]

FIG. 1 is a sectional view of the facsimile device of this embodiment, FIG. 2(A) is a basic control block diagram of the facsimile device of this embodiment, and FIG. 2(B) is the basics of the MPU 23 of FIG. 2(A). A diagram showing the functions, Figure 3 (A) is the ROMI of Figure 2 (A)
A diagram showing the structure of MH code data in I, Figure 3 (B)
, (C) is a flowchart of conversion from run-length code to MR code, and Figures 4 and 5 are EO in RAM9.
FIG. 6 is a flowchart of conversion from MH code to run-length code, FIG. 7 is a diagram showing a search example when converting MH code to run-length code, and FIG. 8 (A) 8(B) is a diagram showing an example of a search when converting an MR code into a run-length code, and FIG.
A diagram showing a processing flowchart of the PU 23, FIG.
) are flowcharts of conversion from run-length code to raw data, FIGS. 8(D) and (E). (F) is an explanatory diagram of the conversion of the number of dots from B4 to A4, Figures 9 (a), (b), and (c) are flowcharts for determining the 14 operation modes of the MPU 23,
11 is a diagram showing the flow of image data in mode M1, FIG. 12 (A) is a diagram showing the flow of image data in mode M2, and FIG. CG
Data in MH code, image data in MR code in RAM
Flowchart diagram for storing data in 9, Figure 13 is mode M
14(A) is a diagram showing the flow of image data in mode M4, and FIG. 14(B) shows the flow of image data in mode M4.
-4, Figure 14 (C) is a diagram showing the page end subroutine, Figure 15 is the mode M5.
16 is a diagram showing the flow of image data in mode M6, and FIG. 17 is a diagram showing the flow of image data in mode M7. M
Figure 18 shows the flow of image data in mode M9.
．． A diagram showing the flow of image data in MIO, FIG. 19 is a diagram showing the flow of image data in mode Mll, FIG. 20 is a diagram showing the flow of image data in mode M12, and FIG. 21 is a diagram showing the flow of image data in mode Mll.
Figure 22 is a diagram showing the configuration of a file management label attached to the top of a page when image data is stored in M9.
Flowchart 1 to automatically clear image data in AM9
- Figure 23 (A) is a diagram showing the flow of image data in mode M13, Figure 23 (B) is a diagram showing the time management subroutine, and Figure 24 is a diagram showing the flow of image data in mode M14. be. In the figure, 1 is a reading unit, 3, 5.7 are RAMs, 9 is FIFORAM used as an image memory, 23 is an MPU, and 25 is a CG.

Claims (3)

[Claims] (1) A reading means for reading an image and converting it into an image signal, an encoding means for compressing and encoding the image signal, a storage means for storing encoded image data, and a transmission means for transmitting the image data in the storage means. In the image transmitting device, the storage means stores a line end code indicating the end of one line of the image data in a predetermined format, and the transmitting means stores the line end code in a format different from the predetermined format. An image transmitting device characterized in that it transmits together with image data. (2) The image transmitting device according to claim 1, wherein the codeword length of the line end code stored in the storage means is set to be longer than the maximum codeword length of the image data. . (3) The image transmitting apparatus according to claim 1, wherein the code word length of the line end code stored in the storage means is longer than the number of parallel processing bits of the transmitting means.