CN1276650A

CN1276650A - Scanned image real-time compression/decompression method

Info

Publication number: CN1276650A
Application number: CN 99106976
Authority: CN
Inventors: 陈皇村
Original assignee: Mustek Systems Inc
Current assignee: Hongyou Science & Technology Co Ltd
Priority date: 1999-06-03
Filing date: 1999-06-03
Publication date: 2000-12-13
Anticipated expiration: 2019-06-03
Also published as: CN1107381C

Abstract

The method of the invention sets each bit group format to have an identification code and a counting code when compressing and coding. The read data is processed by bit planes, which are distinguished into a high bit plane and a low bit plane. When reading the output data of the analog/digital conversion device, 63 bits at most can be read per bit group. Each bit group identifier is the coding property of the bit. The count code section accumulates successive code numbers of the same property. After encoding, the byte is output to the random access memory for the host to access. When decompressing, the bit group identification code is inquired to know whether the code is high/low bit or uncoded data. Then, the codes are sequentially filled into the corresponding positions in the buffer according to the number of the counting codes.

Description

Real-time compression/decompression method for scanned image

The present invention relates to a kind of scan image compressing/decompressing method, be particularly related to a kind of utilization position (bit) and be the scan image compressing/decompressing method of reading unit, repeat digital possibility to improve, and be applicable to the analog/digital conversion device of future 12,14 or more high-order parsing.

The main purpose of image compression/decompression technology is not outer in the speed of quickening image transfer.Being applied to the compression/de-compression technology of scanner (scanner), is the transmission that is applied between scanner and the main frame, is sent to the speed of main frame to speed image.After the image of scanning converts numerical data to through the analog/digital conversion device, just can begin to carry out the compression of image.View data after the compression is sent to a random access memory, for the main frame access.Main frame just begins to carry out decompression, treats just to may be output to phosphor screen after entire image is finished decompression, further processes for the user.

The processing of known scan image is that unit handles with 8 (bit) all, that is only reads 8 positions at every turn when reading the view data that the analog/digital conversion device exported, comes compression/de-compression at the characteristic of the hyte that is read again.This read method is not more had elasticity, and is increased to 12,14 or when more high-order, this reads mode and can't be suitable for from original 8 when the parsing of analog/digital conversion device.And, serve as to handle unit all, because the relation that distortion is handled is easy to generate the phenomenon as mosaic with 8 at every turn.Very and, 8 data volume is little, its repeatability is also less relatively, the efficient of compression is also relatively poor.

Main purpose of the present invention is to propose a kind of image compression/decompression method that can not produce mosaic phenomenon, and seeks can make compression the time to such an extent that repeat digital possibility and improve, and then reaches the purpose of Fast Compression/decompression.

Another object of the present invention is to propose a kind of image compression/decompression method that is not subject to the analyzing efficiency of analog/digital conversion device, so that the method for the compression/de-compression of scan image is applicable to the analog/digital conversion device of future 12,14 or more high-order parsing.

According to above-mentioned purpose, compression method of the present invention is when coding, and the form of setting each hyte (byte) earlier reaches count code (6 bits) two parts for having identification code (2 bits).And with the data that read with bit plane, distinguish high bit plane and low bit plane is handled respectively.When reading the data that the analog/digital conversion device exported, each hyte can read 63 (2 at most ⁶-1) individual position.The identification code of each hyte is represented this coding character, as high/low position or uncoded data.The figure place that the count code part then adds up the coding of continuous same nature.After finishing, coding just hyte is exported to a random access memory for the main frame access.

When decompressing, main frame is the data in the access random access memory just.Simultaneously, and a buffer that can hold 63 hytes is at least arranged, for the data after temporary the decompression.During decompression, as long as the identification code of inquiry hyte, what just know its coding representative is high/low position or uncoded data.And then according to the number of its count code, insert in regular turn in the buffer in the corresponding position, so just can finish the program of decompression.After finishing, just the view data of exportable whole buffer is done further to handle for the user.

Figure 1A is the hyte form of coding of the present invention.

Figure 1B is the identification code kind of the hyte of coding of the present invention.

Fig. 2 A is the processing mode schematic diagram of bit plane.

Fig. 2 B is another schematic diagram of the processing mode of bit plane.

Fig. 3 is the flow chart of compression of the present invention.

Fig. 4 is the flow chart of decompression of the present invention.

Now the present invention is described in detail in conjunction with the accompanying drawings.

For break the mode that reads of habitual view data at all, the present invention is the form of the hyte (byte) of definition coding earlier, shown in Figure 1A.In a hyte, identification code 11 is represented in preceding two positions, and count code 12 is then represented in back 6 positions.Identification code 11 has two positions, therefore 4 kinds of combinations can be arranged, shown in Figure 1B.Wherein, " 00 " represents the data that are all a high position coded in this hyte, i.e. " 1 "." 01 " is represented codedly in this hyte to be the data of low level, i.e. " 0 "." 10 " represent that the data in this hyte all are still uncoded data, i.e. initial data.For example, when the repeatability of view data is very low, just must can directly not export through coding." 11 " are then represented standby, and are still undefined.Certainly, these four kinds of possible combinations and definition can vary with each individual, as long as can represent a high position respectively, and low level, and coding does not wait three kinds to get final product.

The hyte of every coding has 8, after the identification code of removal front two, and 6 last positions, the number of continuous same nature coding when can be used to record coding.Therefore, 6 positions can write down 63 (2 at most ⁶-1) position of individual coding.For example, " 00111111 " just represents to have the view data of continuous 63 high positions (" 1 ").And " 01001111 " just represents to have the view data of continuous 30 low levels (" 0 ").

In addition, the present invention adopts the idea of " bit plane section " to handle the view data that each reads.According to the technology of " bit plane section ", be with 8 bit representations as if each pixel in the image, just the plane that image can be used as by eight 1 constituted, shown in Fig. 2 A.Its scope by least important bit plane 0 to most important bit plane 7.According to 8 binary bit group, plane 0 comprises forms two-symbol all lowest orders of image pixel, and 7 on plane comprises all and forms two-symbol all highest orders of image pixel.Shown in Fig. 2 B, each position of each hyte of reading in will lowly by height be handled respectively.Just, the order of processing position is [A7] [B7] [C7] ... [A6] [B6] [C6] ... [A5] [B5] [C5] ... [A4] [B4] [C4] ... [A3] [B3] [C3] ... [A2] [B2] [C2] ... [A1] [B1] [C1] ... [A0] [B0] [C0] ....

Principle according to this, image is if present with wherein a kind of bit-plane image, and then 8 bit planes will demonstrate diverse image.The compressibility of high bit plane is higher, and the data repeatability of low bit plane is less, and therefore compression is difficult.For the bit-plane data of last three low levels, can encode and directly transmission.Or after the data of transmitting high bit plane, replenish the data of low bit plane again in random number mode, with the distortion rate of data behind the reduction decompress(ion).This is the technology that those skilled in the art used always, and its purpose is only being compressed at the big image of compression efficiency.

According to the above-mentioned coded format and the processing method of bit plane, the data that method of the present invention reads at every turn mostly are most 63 hytes, because there are 6 positions can write down 63 data at most, and every data are used as 8 bit planes and are handled, so each block of handling just has 63 hytes.

The detailed step of compression method of the present invention is as shown in Figure 3: 301: read the data that the analog/digital conversion device is exported; 302: because treatment of picture is to handle item by item in regular turn, therefore with the image of each bar scanning,

With 63 hytes is a block, is divided into several and handles block; 303: handle the data of the block that is read in regular turn, its handling procedure is as follows: 3031: read 6 positions that the analog/digital conversion device is exported in regular turn; 3032: the attribute of judging these 6 positions? when these 6 positions are all " 1 ", execution in step 3033; When

When these 6 positions are all " 0 ", execution in step 3034; During other situation, execution in step 3035; 3033: the identification code of setting this hyte is " 00 ", and this count code is made as two-symbol " 11 ",

And execution in step 3036.3034: the identification code of setting this hyte is " 01 ", and this count code is made as two-symbol " 11 ",

And execution in step 3036.3035: the identification code of setting this hyte is " 10 ", and execution in step 3038; 3036: judge that whether next bit is with value, or six positions of this count code are all " 1 "? if,

Just execution in step 3037; Otherwise just execution in step 3038.3037: with this count code progression 1, and repeated execution of steps 3036.3038: export this hyte to one storage device.3039: judge whether to finish the processing of this block? if then execution in step 304; Otherwise, carry out

Step 3031; 304: judge whether to finish the processing of all blocks? if then execution in step 305; Otherwise execution in step 303.305: finish.

In step 3031,, can earlier binary code be converted to Ge Lei sign indicating number (gray code) for increasing the compressibility of data.The Ge Lei sign indicating number is that the computing of XOR is carried out in adjacent position.Because the Ge Lei sign indicating number has unique character, promptly, between the code word that links to each other a position difference is only arranged through after the conversion of Ge Lei sign indicating number.Therefore, utilize this characteristic, even GTG has very little variation, also can show at all bit planes, it is a sheet of complete black or complete white can not making picture.

Data after the compression are temporarily stored in the random access memory, handle for decompressing device.The method that decompresses is to carry out with respect to compression method, and its method is as follows, and decompression needs the data after the buffer stores of at least 63 hytes decompresses.The memory of buffer is arranged the arrangement that is equal to the preceding view data of compression.The detailed step of decompression of the present invention is as follows: 401: read coded data in regular turn, read a hyte at every turn.402:, judge the attribute of this hyte by the identification code of this hyte.403: when the identification code of this hyte is " 00 ",, comply with according to the number that counter write down of this hyte

Preface is inserted equivalent " 1 " corresponding position in buffer; Promptly according to

[A7] [B7] [C7] ... [A6] [B6] [C6] ... [A0] [B0] [C0] ... wait order, open by a high position to low level

Beginning is inserted " 1 " of equivalent.404: when the identification code of this hyte is " 01 ",, comply with according to the number that counter write down of this hyte

Beginning is inserted " 0 " of equivalent.405: when the identification code of this hyte is " 10 ", then, deposit in regular turn with the data of the count code of this hyte

Go into corresponding position in the buffer.Because this hyte is not encoded as yet, therefore as long as with count code

Data insert relative position and get final product.For example, if this hyte is " 10101010 ", and

Represent the 5th data, then in regular turn " 101010 " are inserted respectively

Among [A5] [B5] [C5] [D5] [E5] [F5].406: judge whether buffer fills up? if, execution in step 407; Otherwise, execution in step 401.407: the data in the output buffer.408: judge whether to finish the decompression of all data? if, execution in step 409; Otherwise, carry out

Step 401.409: finish.

So after the view data reduction of each block, just with the input of the data in buffer main frame, until the conversion of finishing all blocks.Wherein, if original binary code has been converted to the Ge Lei sign indicating number, when execution in step 407, be the Ge Lei sign indicating number owing to what decompress, must earlier the Ge Lei sign indicating number be converted to binary code, could be with the input of the data in buffer computer.Compression and the step that decompresses can be carried out simultaneously, must be unit with 63 hytes just.After condensing routine was 63 hytes with image encoding, is sent to storage device then, gunzip just can begin to carry out.Because compression the time is to be unit with 63 hytes, therefore, gunzip must be waited until that 63 hytes are all encoded and just can carry out after finishing.

Method of the present invention can software implementation, also program making can be become special integrated chip (ASIC), comes the compression and the decompression of carries out image.In sum, compressing/decompressing method of the present invention is a unit with position (bit) still when reading the data that the analog/digital conversion device exported, but not be that unit handles with hyte (byte), therefore not only has digital equable advantage, and seek easily with the value number, so the compression efficiency of image is higher.

In addition, compressing/decompressing method of the present invention is aspect image impairment (loss), because of taking other processing at one or two position, for example 255 rank are reduced to 128 rank or 64 rank, make image passivation etc., the user also is not easy to discover, and its change is comparatively soft, can not produce the phenomenon of well-known category like mosaic.Very and, compression method of the present invention can be selected the transmission of low level (low bit), transmits with a line (oneline) sampling, or rebuilds with the idea of responsibility cycle period (duty cycle), so the transmission quantity of visual image and deciding, and has very much elasticity.

The above only is preferred embodiment of the present invention, and has reached practical effect widely, and all equivalent variations and modifications of being done according to claim scope of the present invention all belong in the scope that claim of the present invention contains.

Claims

1. A scanned image real-time compression method, comprising steps:

Set a group, the group contains the identification code, and counting code;

Read the data output by the analog/digital conversion device;

Dividing the output data into a plurality of processing blocks;

Process the data of the read blocks sequentially, including steps:

(a) sequentially read the 6 bits output by the analog/digital conversion device;

(b) determine the attributes of the six bits, and set the identification code of the bit group according to the attributes of the six bits;

(c) When the identification code of the group of bytes represents the high or low position, set the count code of the group of bytes as "11" of the binary code, and perform step (e);

(d) When the identification code of the bit group is not high or low, directly input the six counting codes in the bit group;

(e) The count code of the bit group accumulates consecutive digits with the same value until all six digits of the count code are "1", or the next digit is not the same value;

(f) When the six bits of the count code are all "1", or when the next bit is not the same value, output the group of bits;

(g) Repeat steps (a) to (f) until the encoding of all image data is completed.

2. The method as claimed in claim 1, wherein said step (a) further comprises the steps of:

(h) When reading the 6 bits output by the analog/digital conversion device, converting the 6 bits into a Gray code.

3. The method of claim 1, wherein the identification code is two digits, and the counting code is six digits.

4. A scanned image real-time decompression method, comprising steps:

Read the encoded byte data sequentially, one bit at a time;

According to the identification code of the read group of bytes, the attribute of the group of bytes is judged;

When the identification code of the read group of bytes indicates a high bit, according to the number recorded by the count code of the group of bytes, fill in the corresponding position in the buffer with an equal amount of "1" in sequence;

When the identification code of the read group of bytes indicates the low bit, according to the number recorded by the count code of the group of bytes, fill in an equal amount of "0" in the corresponding position in the buffer in sequence;

When the identification code of the read group of bytes is not low or high, sequentially store the count code data of the group of bytes into the corresponding position in the buffer; and

When the buffer is full, the data in the buffer is output.

5. The method of claim 4, further comprising the steps of:

When outputting the data in the buffer, convert the Gray code into a binary code.

6. The method of claim 4, wherein said buffer can accommodate at least 63 bursts.

7. A scanned image real-time compression/decompression method, comprising steps:

(a) set a group, the group contains the identification code, and counting code;

(b) read the data output by the analog/digital conversion device;

(c) dividing the output data into a plurality of processing blocks;

(d) sequentially read the 6 bits output by the analog/digital conversion device;

(e) determine the attributes of the six bits, and set the identification code of the bit group according to the attributes of the six bits;

(f) When the identification code of this group of digits indicates high or low digits, set the counting code of this group of digits to "11" in binary code, and accumulate consecutive and identical digits in the counting code of this group of digits, Up to the six digits of the count code are all "1", or the next digit is not the same value;

(g) When the identification code of the bit group is not high or low, directly input the counting codes of the six original data in the bit group;

(h) When the six digits of the count code of the bit group are all "1", or when the next bit is not the same value, output the bit group;

(i) repeating steps (d) to (h) until the encoding of all image data is completed;

(j) read the coded byte group data sequentially, and read one group at a time;

(k) judge the attribute of the bit group according to the identification code of the bit group read;

(l) When the identification code of the read group of bytes indicates a high bit, according to the number recorded by the count code of the group of bytes, fill in the corresponding position of the buffer with an equal amount of "1" in sequence;

(m) When the identification code of the read group of bytes indicates the lower bit, according to the number recorded by the count code of the group of bytes, fill in the corresponding position of the buffer with an equal amount of "0" in sequence;

(n) When the identification code of the read group of bytes is not low or high, sequentially store the count code data of the group of bytes into the relative position in the buffer; and

(o) When the buffer is full, output the data in the buffer.

8. The method of claim 7, wherein said step (b) further comprises the steps of:

(p) When reading the 6 bits output by the analog/digital conversion device, converting the 6 bits into Gray codes.

9. The method of claim 7, wherein said step (o) further comprises the steps of:

(q) converting the Gray code into a binary code when outputting the data in the buffer.

10. The method of claim 7, wherein the identification code is two digits, and the counting code is six digits.

11. The method of claim 7, wherein said buffer can accommodate at least 63 bursts.

12. The method of claim 7, wherein said steps (b) to (i) and said steps (j) to (o) are performed simultaneously.