JPH0666865B2 - Image transmission method - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to an image transmitting method of reading a document image to obtain an image signal, compression-coding the obtained image signal, and transmitting the compressed signal.

<Background of the Invention> In a facsimile machine, a minimum transmission time is set between a transmission side and a reception side. The minimum transmission time is defined as the time required for reading and encoding one line of the original on the transmitting side and the time required for recording one line on the receiving side. If the image signal of one line is transmitted in a time shorter than the minimum transmission time, recording will not be in time on the receiving side.

Therefore, as shown in Fig. 1, in the G3 mode facsimile equipment according to the CCITT, T30 recommendation, when the transmission time of the encoded information of one line is less than the minimum transmission time, all the dummy bits are added and transmitted. It was

On the other hand, the applicant of the present invention has proposed a transmission method shown in FIG. 1 as a file bitless transmission mode in Japanese Patent Application No. 58-209826.
Proposed in the issue. This is set so that the file bit is not transmitted if the total transmission time of a plurality of lines is equal to or longer than the minimum transmission time of the number of lines. By setting in this way, the amount of the file bit can be reduced only in the portion indicated by A in the figure, and the transmission time can be shortened.

On the other hand, the amount of file bits in the file bitless transmission mode was calculated as follows.

First, the total capacity of the reception buffer that temporarily stores the reception signal on the reception side is detected, and the data amount is added every time one line of image data is transmitted, and when the added data amount exceeds the capacity of the reception buffer, When the amount of added data is larger than the number of bits corresponding to the minimum transmission time for the number of lines, the file bit is not transmitted. When the amount of added data is smaller than the number of bits, the bit number of the difference is added to the image data of one line before as a file bit and sent.

In this way, the amount of added data is compared with the total capacity of the reception buffer. However, during transmission of data, the data in the reception buffer is read and recorded, so even if the amount of added data exceeds the total capacity of the reception buffer, there is actually an empty area in the reception buffer. Should be. That is, the reception buffer is not being used effectively.

Since the capacity of the buffer memory is greatly related to the device cost, it is desirable to effectively use the buffer memory having the smallest capacity.

<Objective> In view of the above points, the present invention has an object to provide an image transmitting method capable of effectively using a buffer memory on the receiving side.

More specifically, in the image transmitting method of compressing and encoding the original image signal and transmitting the same via the modem, the encoded image is stored in order to estimate the free space of the buffer for temporarily storing the received image signal on the receiving side. Using a read address of the memory and a virtual address indicating a processing end position on the receiving side in the memory, the read address is stepped in units of the number of bits supplied to the modem each time an encoded image is supplied to the modem. It is an object of the present invention to provide an image transmitting method characterized in that the virtual address is stepped on a line-by-line basis at each time required for processing one line of encoded image on the receiving side.

(Embodiment) An embodiment of the facsimile apparatus for realizing the present invention will be described in detail below.

(Mechanical System) FIG. 2 (A) shows a sectional view of the facsimile machine. In the figure, 41 is a CCD solid line image sensor, 42 is an imaging lens, 43 is a mirror, 44 is a document illumination lamp, 45 is a document feeding roller, 46 is a document discharging roller, 47 is a document feeding tray,
Reference numeral 31 is a document detection sensor that detects the presence or absence of a document on the paper feed tray.

Further, 34 is a roll paper storage cover, 35 is roll paper, 36 is an original and recording paper discharge tray, 37 is a cutter, 38 is a roll paper discharging roller, 39 is a roll paper conveying roller, 40 is a recording head, and 33 is a recording head. A roll paper cover sensor that detects opening / closing of the cover 34.

In the figure, at the time of reading the original, the original on the original feeding tray 47 is conveyed by rollers 45 and 46. The original is illuminated by the lamp 44 at the reading position P, and the reflected light is reflected by the mirror 43 and the lens 42.
An image is formed on the image sensor 41 via the, and the image sensor 41 converts the image into an electric signal.

On the other hand, at the time of recording, the roll paper 35 is nipped by the roller 39 and the head 40 and conveyed, and at the same time, the head 40 is put on the heat-sensitive roll paper 35.
An image is formed by. When the recording for one page is completed, the roll paper 35 is cut by the cutter 37 and discharged onto the paper discharge tray 36 by the roller 38.

(Basic Block) FIG. 2 (B) is a basic control block diagram of the facsimile machine of this embodiment. In the figure, 1 is a reading unit for reading an original image and converting it into an electric image signal, 3, 5 and 7 are one of the modes, a random access memory (hereinafter referred to as RAM) which functions as a buffer for temporarily storing the image signal, and 9 is a code. First-in first-out RAM that stores the converted image signal
(Hereinafter referred to as FIFORAM), 11 is a read-only memory (hereinafter referred to as ROM) that stores the MPU23 operation program, 13 is RAM that stores flags and data necessary for the MPU operation, and 15 is an operation that has an input key, a display, etc. Reference numeral 17 is a recording unit for recording a copy image, a received image and management data on a thermal paper, 19 is a modem for modulating transmission data and demodulating the reception data, 20 is a telephone, 21 is a communication line 22 and a modem 19 or A network control unit (hereinafter referred to as NCU) that controls connection to the telephone 20 and 23 is an MPU that controls the entire system. In this embodiment, as the MPU 23, a 16-bit data bus 24 and an Intel 8086 capable of directly accessing a memory space of up to 4 megabytes are used.

There are six basic functions of the MPU23 as shown in Fig. 2 (C). Each function will be described in the flow of image data.

(Transmission Operation) The flow of image data during the transmission operation will be described with reference to FIG. In this example, an MH code transmission sequence is shown.

The reading unit 1 converts the image data for one line into a run length code RL and writes the run length code RL in the RAM 3 in response to a read command from the MPU 23. Then, the MPU 23 transfers the data in the RAM 3 to the two line buffers RAM 5 and RAM 7 as they are, alternately by one line, and encodes the run length code RL read from the two line buffers into the MH code and writes it into the FIFORAM 9. Then, in response to the data request interrupt from the modem 19, the MPU 23 transfers the MH code from the FIFORAM 9 byte by byte to the modem. At this time, if necessary, a file bit is inserted. The insertion of the file bit will be described later in detail. The 1-byte data transferred from the FIFORAM 9 to the modem 19 is transmitted to the receiving side via the telephone line via the NCU 21.

Here, except for the case of the reading unit 1-> RAM3 and the modem 19-> NCU21 in the figure, all other data transfer is performed via the bus 24 of the MPU23.

In the data request interrupt from the modem 19, the interrupt interval changes depending on the transmission rate. Data transfer is done in bytes, so 8/9600 at 9600 bps
= An interrupt occurs every 0.83 × 10 ^-3 sec.

Further, when the data transfer from RAM3 to RAM5, RAM7 is completed, the MPU 23 outputs a read command to the reading unit. While the MPU 23 is performing the encoding process ENC and the interrupt process, the reading unit 1 reads the document and converts the raw data to the run length data.

(Reception Operation) The reception operation of the MH code signal will be described with reference to FIG. MP
When the U23 receives the MH code through the NCU21 and the modem 19 from the selection, it deletes the file bit first and transfers it to the RAM9 as it is in the MH code. Then, the MH code is sequentially exposed from the RAM 9, converted into the line length code RL, transferred to the recording unit 17, and recorded.

(Management of FIFORAM9) Next, a management method of the FIFORAM9 during transmission will be described. FI
In this embodiment, three kinds of address pointers are used for management of FORAM9.

The first pointer is called the encoded pointer EP. EP indicates the final storage address of the encoded data for which the encoding process has been completed. The EP is incremented by 1 byte. This step is managed by a main routine, which will be described later, performed every time one byte is encoded.

The second pointer is called the modem pointer MP. MP indicates the final storage address of the encoded data that has been sent to the modem 19 by the modem interrupt. The MP is managed by a modem interrupt routine which will be described later. Since one byte of data is sent out by one interrupt, the MP is incremented by one byte.

The third pointer is called a simulation pointer SP.
SP indicates the final storage address of the encoded data that has already been recorded on the receiving side. SP is incremented by one line of data. This step is executed at every minimum transmission time of the receiver and stepped to the next memory address of the EOL signal. The SP is managed by a timer interrupt routine described later.

The minimum transmission time of the receiver is notified to the transmission side by a procedure signal exchanged before the image data is transmitted.

The relationships among the pointers EP, MP, SP are shown in FIGS. 5 (A) and 5 (B). In FIG. 5 (A), the lower side is the address (00
0), the upper side is the address (FFF). Pointer EP, MP, SP
Each goes from the bottom to the top, and when it reaches the top, it goes back to the bottom again.
Here, the area 101 is an area for storing data which has been encoded and transferred and whose recording processing has been completed on the receiving side. Therefore, the data in the area 101 is already unnecessary.

Further, the data in the area 102 is data that has been encoded and sent to the modem 19, but has not been processed by the receiving side, and is data being transmitted or stored in the buffer on the receiving side. . Therefore, the data in the area 102 is already unnecessary data.

Further, the data in the area 103 is data that has only been encoded, and is currently data that has not been transferred to the modem 19.

Finally, area 104 is an area that still contains no encoded data.

EP, MP and SP must meet the following conditions C1-C3.

C1 ... MP can't overtake EP. When the MP catches up with the EP, it sends a file bit, executes only encoding, and stops sending image data.

C2 ... SP cannot overtake MP. Since the receiving side cannot record data that has not been transmitted, the SP is stopped when it catches up with the MP.

C3 ... EP cannot overtake SP. EP is the top address (FF
This is because the data in the area 102 is rewritten when the SP is overtaken while the coded data is accumulated from the lower end again after reaching F), and the step destination of the SP becomes unknown. Therefore, when the EP catches up with the SP, the encoding process is suspended.

The buffer 9 is managed so as to satisfy the above three conditions.

Consider here an example in which three pointers EP, MP, and SP overlap. An example is shown in FIGS. 6 (A), (B), and (C).

The situation as shown in FIG. 6 (A) occurs when the pattern of the document is complicated and requires a lot of time for encoding. In this case, the modem 19 waits for the encoded data to be ready while transmitting the file bit. At this time, the receiving side is also waiting for the next line of data.

FIG. 6 (B) shows a case where the encoding and the processing on the receiving side proceed quickly and the transmission cannot catch up. In this case, the transmitter has interrupted reading and encoding and is waiting for transmission processing. On the other hand, the receiving side is also waiting for the next line of data.

FIG. 6 (C) shows a case where the coding and transmission proceed rapidly and the processing on the receiving side cannot catch up. At this time, the transmission side suspends encoding and transmission until the processing on the reception side proceeds. Actually, the EP may pass the SP, but the encoding is interrupted for the reason explained in the condition C3.

The MPU 23 that realizes the management method of the buffer 9 described above
The control operation of will be described below.

(Main Routine) FIG. 7 shows a control flow chart of the main routine.

First, in step ST1, initial setting is performed. Here, EL indicates the number of EOL signals in the area 103 in FIG. 5A, that is, the number of untransmitted coded lines. Further, ML indicates the number of EOL signals in the area 102, that is, the number of unprocessed lines on the receiving side.

When the initial setting is completed, encoding is started in ST2. And S
When one byte of data is prepared at T3, it is determined at ST4 whether EP and SP are equal. If EP and SP are equal, EP cannot pass SP as described in condition C3, and therefore encoding and storage of encoded data are interrupted. If EP and SP are not equal, one byte of encoded data is stored in the address EP of RAM 9 in step ST5. Then, EP is incremented (ST6). Step ST2 until encoding of one line is completed
~ Repeat ST6. EO when line end is detected in ST7
The L signal is stored in RAM9.

Here, the form of the EOL stored in the RAM 9 will be described.

The transmitted EOL consists of 11 consecutive "0s" and one "1".

The MPU23 adds this E to the image data for each line end detection.
The OL is added and sent, but when adding this EOL,
The transmission time of the transmission line is calculated, and if it is less than the minimum transmission time, the value EO with the file bit inserted.
L is added. The detection of EOL which is the line end signal when reading from RAM9 becomes an important issue. Therefore, this device uses the following methods to simplify EOL detection and EOL transmission when reading from RAM9.

First, as a basic idea of handling EOL (1) EOL is added when writing to RAM9.

(2) EOL detection when reading from RAM9 is 0 for 2 consecutive bytes
Done in.

(3) When sending data from RAM9, 0 for 2 consecutive bytes
Of these, 0 of the 2nd byte is not transmitted.

There are three points. The two cases will be described below, depending on the case.

FIG. 8 shows the storage form of the data and EOL in the RAM 9 when "1" in the final data of the line exists in the final byte. In the figure, DT of the last byte A is image data. The byte A is filled with 0 after the data DT, the bytes B and C are all set to 0, and 1 is inserted at the D byte. However, depending on the number of 0s inserted in the byte A, the number of 0s before the 1 in the D-th byte is changed as shown in the table below.

In this way, 11 0s can be secured even if EOL 0s in the memory are sent after removing 1 byte.

Next, the storage form of the data and EOL in RAM9 when the "1" in the final data of the line does not exist in the final byte is
Shown in the figure. Data contained in the last byte A as shown in the figure
When DT is all 0s, all 0s are inserted in the rest of byte A, and 0s are also inserted in the next byte B. Then, in the byte C, the number n of 0s inserted in the byte is subtracted from 1 and 0 is inserted, and then 1 is inserted.

Since there is no last consecutive 0 or more 4 in the MH code, the number of 0s inserted in the A byte is not considered to be 4 or less.

Further, in consideration of white line skip transmission, in the case of S in which all the data of one line is all white, 0 of the second byte is distinguished as "01" (hexadecimal display) as a criterion of all white data.

By writing to FIFORAM9 with the above format, when reading from RAM9 (during modem interrupt processing)
EOL detection is 2 bytes consecutive 0 or 1 byte 0 and “01”
(Hex) makes it easy. Further, when the read data is transmitted, the EOL can be easily transmitted by deleting the 0 (or "01") of the second byte.
EOL can be sent without deleting the second byte 0, but by deleting it, unnecessary data can be prevented from being sent and the transmission time can be shortened. Returning to the main routine again, description will be made.

At the same time as setting the EOL at step ST8, the EP is incremented by the number of bytes of the EOL. Since the encoded data of one line is stored, EL is incremented in ST9. The above operation of ST2 to ST9 is repeated until the transmission of the image data for one page is completed. When the end of storing one page of image data is detected in step ST10, ST11
Then, the RTC data is set in the RAM 9 and further set in the RTC flag RAM 13 indicating that the RTC can be transmitted. Then, after that, in ST12, it is judged that EF = 0, that is, the data of one page is sent out, and the operation is ended.

(Modem interrupt routine) The modem interrupt routine will be described with reference to FIG. The modem 19 generates an interrupt signal every time one byte of data is transmitted. The processing of the main routine in which an interrupt occurs is suspended, and the modem interrupt routine is entered.

In ST21, it is judged whether EL = 0, that is, whether or not there is untransmitted coded data. If untransmitted coded data exists, 1-byte data of the address MP is read from the RAM 9 in ST22. Then, in ST23, it is determined whether or not the data is "0000 0000" at the second byte which is continuous for 2 bytes. In the case of "0000 0000" in the second byte, it is clearly the line end, so the operation proceeds to step ST26. If not the line end
In ST24, the data is sent to the receiving side via the NCU, MP is incremented, and the process returns to the main routine.

On the other hand, when ST23 detects "0000 0000" in the second byte, it is determined in ST26 whether the number of untransmitted coded lines EL is 1 or not. In case of 2 or more, increment MP in ST27 and 2
Do not send the "0000 0000" at the byte, and set the next 1 in ST28.
Read the byte data and send the data in ST29. Then, again in ST30, MP is incremented. Since the transmission of one line is completed, EL is decremented in ST31, ML is incremented in ST32, and the process returns to the main routine. On the other hand, if EL = 1 in ST26, it means that the MP has caught up with the EP. This is because the encoding can be sent rather than the electric transmission.
This is the case where the encoding of all lines is completed and the last line is being transmitted. ST33 is the number of unprocessed lines on the receiving side ML
Is judged to be 0 and the file bit is transmitted while there is unprocessed data. Also, if ML is 0 and the RTC flag is not set, it is not the last line, so a file bit is sent in ST37. When ML = 0 and the RTC flag is also set, the receiving side has finished recording all lines, so EL is set to 0 and the first data of the RTC signal indicating that the data transmission of one page is finished is sent to the receiving side. , Return to the main routine.

On the other hand, if EL = 0 in ST21, check the RTC flag and set R
If the TC flag is set, the remaining RTC data is sent out and the process returns to the main routine. On the other hand, if the RTC flag is not set, it means that the transmission of all data has not yet been completed, so that the file bit is transmitted and the process returns to the main routine.

(Timer Interrupt Routine) Next, the management of the simulation pointer SP by the timer interrupt routine will be described with reference to FIG. The timer interrupt signal is generated at the minimum transmission time interval on the receiving side, and the timer interrupt routine is executed. First, in ST41, it is checked whether or not the number of unprocessed lines ML on the receiving side is zero. If ML is 0, the receiving side is waiting for data from the transmitting side, and therefore the process directly returns to the main routine. If ML is not 0, it is judged that the recording of one line is completed at that time, and the storage address of the EOL next to the storage address of the current EOL of the pointer SP is searched (ST42). Then, in ST43, the storage address of the next EOL is set to SP. In ST44, decrement one ML and return to the main routine.

Incidentally, since the simulation pointer SP is estimated by the transmitting side, a quantization error occurs. For example, if a line in which the transmission time of one line of image data is longer than the minimum transmission time of the receiver continues, the pointer array as shown in FIG. 12 is obtained. Figure 12 shows that SP is at the end of line N-1 and MP
Is in the middle of line N. At this time, SP is not incremented because ML = 0 even if a timer interrupt is applied. On the other hand, even if the MP finishes the transmission of the line N in 1 ms from this timer interrupt, the SP is not stepped to the next interrupt (for example, 5 msec interval).
1 = 4msec SP will be delayed from the actual processing on the receiving side. However, this quantization error may be delayed but not advanced, so that no data loss occurs.

(If the sending side and receiving side have different buffer sizes) For the above buffer management, the receiving side's buffer size capacity is
We have described the same case as the sender. Even when the buffer size on the receiving side is larger than that on the receiving side, the same method as described above may be used. This is because the amount of unprocessed data on the receiving side exceeds the capacity of the transmission buffer, which cannot be achieved by the above buffer management.

However, if the buffer size on the receiving side is smaller than that on the transmitting side, the amount of unprocessed data on the receiving side, that is, the pointer MP
It is necessary to control so that the amount of data between SP and SP does not exceed the buffer size on the receiving side.

Therefore, one pointer was added as shown in FIG. This pointer is called a simulation upper limit pointer SP '.
SP 'is always placed at a position preceding the simulation pointer SP by 150 minutes on the receiving side buffer capacity. And E
P is controlled so that it cannot overtake SP '. As mentioned earlier M
P can't overtake EP. Therefore, since the MP cannot pass SP ', the buffer capacity between the MP and SP is always smaller than the buffer size on the receiving side, and the buffer on the receiving side does not overflow.

The processing flow chart in this case can be realized by partially modifying the main routine of FIG. 7 and the timer interrupt routine of FIG. First, the main routine is step
Step ST1 'is executed instead of ST1. In ST1 ', instead of setting SP to 0, the storage byte amount X of the buffer on the receiving side is set to SP'. And in ST4 'EP = SP'
Determine whether or not. When EP and SP 'are equal, encoding is interrupted. Others are the same as the main routine described above. Also, the modem interrupt routine is the same as previously described and does not need to be modified.

Next, the timer interrupt routine is ST43 'instead of ST43'.
To execute. In ST43 ', SP is incremented to the next memory address of EOL and X is added to SP to obtain SP'.

(Summary) As described above in detail, in the present embodiment, regardless of the size of the buffer on the receiving side, the processing amount of the image data on the receiving side is calculated using the minimum transmission time notified before the data transmission. It is estimated so that the amount of data that has been sent to the receiving side but not yet processed by the receiving side does not exceed the capacity of the receiving side buffer memory.

Therefore, the memory on the receiving side does not overflow,
Moreover, it is possible to effectively use the already processed empty area of the memory on the receiving side. Further, it becomes possible to significantly reduce the number of times of transmitting the bit bit and the amount of the bit bit by stopping the transmission of the image data on the transmitting side.

Therefore, it is possible to transfer images at high speed with a small amount of memory. Further, since the number of interruptions of reading and recording is reduced, it contributes to improvement of image reading quality and recording quality.

(Other Embodiments) In this embodiment, an example in which the minimum transmission time on the receiving side is constant has been described. However, when the maximum white transmission time is all white pixels on one line, there are black pixels even in part. Even if it is different from the case where it is performed, it can be realized by changing the interval of the timer interrupt according to the image data.

<Effect> As described above, according to the image transmitting method of the present invention, the free space of the receiving buffer on the receiving side is estimated using the read address of the memory for storing the encoded image and the virtual address indicating the processing end position on the receiving side. The read address is incremented by the number of bits supplied to the modem each time the coded image is supplied to the modem, and the virtual address is incremented every time it takes to process one line of the coded image on the receiving side. As you step by line,
The free space can be reliably estimated, and high-speed transmission is possible while preventing the overflow of the receiving buffer.

[Brief description of drawings]

FIG. 1 is an explanatory view of a conventional G3 mode and a file transmission method in a file bitless transmission mode, FIG. 2 (A) is a sectional view of a facsimile machine of this embodiment, and FIG. 2 (B) is FIG. 2A is a basic control block diagram of the apparatus, FIG. 2C is a functional block diagram of the MPU 23 shown in FIG. 2B, and FIG. 3 is a block diagram showing the flow of image data during transmission with the MH code. FIG. 4 is a block diagram showing the flow of image data at the time of reception with the MH code, and FIGS. 5 (A) and 5 (B) show pointers EP, MP, SP for buffer management of this embodiment. Figure showing the relationship, No. 6
6A, 6B, and 6C are pointers EP and M.
FIG. 7 shows a case where P and SP overlap, FIG. 7 shows a main routine, FIGS. 8 and 9 show a storage form of EOL in RAM 9, and FIG. 10 shows a modem interrupt routine.
FIG. 11 is a diagram showing a timer interrupt routine, FIG. 12 is an explanatory diagram of the quantization error of the simulation pointer SP,
FIG. 13 is an explanatory diagram of buffer management when the reception buffer size is smaller than the transmission buffer size. In the figure, 1 is a reading unit, 9 is a FIFORAM, 17 is a recording unit, and 19
Indicates a modem, 21 indicates an NCU, and 23 indicates an MPU.

 ─────────────────────────────────────────────────── ─── Continued Front Page (72) Inventor Yasuhide Ueno 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Ken Ken Ono 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Co., Ltd. (72) Inventor Shigeo Miura 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Yosei Oto 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. ( 72) Inventor Masahashi Takahashi 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc.

Claims (1)

[Claims] 1. A memory for storing coded images in order to estimate free space of a buffer for temporarily storing a received image signal on a receiving side in an image transmitting method for compressing and encoding an original image signal and transmitting the same via a modem. Of the read address and a virtual address indicating the processing end position on the receiving side in the memory, the read address is stepped in units of the number of bits supplied to the modem each time an encoded image is supplied to the modem. An image transmitting method, wherein the virtual address is stepped in line units at each time required for processing one line of encoded image on the receiving side.