JPH03143171A - Method and device for compressing group of digital signals - Google Patents

Abstract

PURPOSE:To obtain a group of compressed digital signals which are efficiently compressed by obtaining the group of compressed digital signals through compression by the use of plural compression sections unequal in the length of a value measuring direction and adding a new compression section to the compression section in low approximating accuracy until the approximating accuracy to a group of original digital signals reaches a prescribed value. CONSTITUTION:A compressing means 12 compresses a group of original digital signals whose values change in one value measuring direction by the use of plural compression sections unequal in the length of a value measuring direction. The approximating accuracy of the group of obtained compressed digital signals to the group of original digital signals is obtained by an approximating accuracy computing means 13. A compression section adding instructing means 14 sends a command for adding a new compression section to the compression section in low the approximating accuracy and recompressing it to the compressing means 12 until the approximating accuracy reaches the prescribed value. Thus, the group of original digital signals can be compressed to the group of compressed digital signals correspondingly changing with good response feature and further, can be efficiently compressed with the small number of compression sections.

Description

[Detailed Description of the Invention] [Industrial Application Field] Kimitsu compresses an original digital signal group whose values change with respect to one measurement direction into a compressed digital signal group with a smaller number of signal GVs. The present invention relates to a digital signal NBY compression method A3 and apparatus.

[Prior Art] In general, in order to improve the efficiency of converting any variable value into a group of original digital signals whose values change with respect to one measurement direction and further utilizing this group of original digital signals, Although the tendency of the value change corresponds to the original digital signal, the number of signals is compressed and converted into a compressed digital signal group that is smaller than the number of signals of the original digital signal.

As an example of compressing such a digital signal g group,
For example, as shown in FIG. 22, there is a case where an arbitrary image 1 is printed on paper 3 as a reproduced image 1a.

To explain further, FIG. 22 shows a case where the image @1 and character 2 shown on the leftmost part are printed as reproduced images 1a and 1q raw characters 2a on paper 3 as shown in the rightmost part. ing. One character 2 is converted into a digital signal 0 by a character data input device 4 such as a keyboard, and is input as character data to the host combination coater 6. The other image 1 is read as image data by an image data input device 5 such as an image scanner and is input to the host combi coater 6. This ill! The image data read from iiB 1 is an original digital signal consisting of a set of digital signals obtained by measuring the density of image 1 for each unit pixel along the scanning direction of the scanning line. If this original digital signal is used as it is for printing image 1, it will contain too much information and place a huge burden on the host computer 6, CPU, etc. of the host computer 6, printer 7, etc. It has to be formed with high performance, making it expensive.

Therefore, we have conventionally created a compressed digital signal g that changes according to changes in the original digital signal and has a small number of letters, and uses this compressed digital signal to edit images 1 and characters 2 by the host computer 6. etc.

The print data created by the host computer 6 is sent to the printer 7, and the printer 7 prints the reproduced image 1a and reproduced characters 2a on the paper 3 based on the received print data.

[Problem to be solved by the invention]

However, the conventional method of compressing the original digital Shingetsu group has the following problems.

That is, the reading of the shading of the image 1 using the image data manually is carried out by repeating measurement in the X direction while scanning the image 1 shown in FIG. 22 in the X direction.

FIG. 23 shows image data consisting of the degree of density of the image 1 along the scanning line, obtained by scanning the image 1 in the X direction by the image data input device 5.

This image data uses the gradation value of a degree measured for each pixel in the X direction of the image a1 as a digital signal value, and displays these digital signal values as a line graph. This is a type of original digital signal group. When compressing this image data, as shown in Fig. 23, the entire scanning direction of image 1 in the Tahara Digital? , ”: QB¥ is compressed into one compressed digital signal.

However, as in the image data shown in FIG. 23, the original digital signal group often includes a high frequency component region whose value changes rapidly and a low frequency component region whose value changes slowly. In order to compress the high frequency component so that it changes responsively in response to the change, the length in the measurement direction, that is, the length in the X direction, of the compression section to be equally divided must be shortened.

However, if the length of the compression section in the X direction is reduced,
This creates a larger number of compressed digital signals than necessary in the low frequency component region, resulting in poor compression efficiency including redundant data, resulting in a large amount of data being required for subsequent control processing in the host computer 6 and printer 7. Take the burden! It will be J. For this reason, in the conventional method described above, it is difficult to determine the number of equal divisions of the entire scanning range, that is, the length of the compressed section in the direction of the diaphragm measurement, to an appropriate value.

The present invention has been made in view of these points, and it is an object of the present invention to convert a group of original digital signals whose values change in one measurement direction into a group of compressed digital signals whose values change responsively to the changes. Moreover, it is an object of the present invention to provide a digital signal group compression method and apparatus that can efficiently compress a group of digital signals in a small number of compression sections.

[Means for solving problems]

In order to achieve the above object, the digitization r, H of the present invention
l! The y-compression method is a digital signal group compression method that compresses an original digital signal station group whose values change in one measurement direction using a plurality of compression sections divided in the measurement direction. Using multiple compression intervals with unequal lengths in the value direction, The original digital signal group is compressed to obtain the compressed digital signal Y, and then the original digital signal q1 of the compressed digital signal q1! The feature is that 50 new compression sections are added to the compression section with the lowest approximation f1 degree until the approximation accuracy for y reaches a predetermined value.

In addition, in order to achieve the above object, the digitization; a compression means for compressing using a plurality of compression intervals having unequal values, an approximation v1 degree derivation means for determining the approximation accuracy of the rf1tA digital signal group obtained by the compression means to the original digital Shintan group, and this approximation. Compression section addition instruction means for issuing a command to the compression means to add a new compression section to the compression section with low approximation accuracy and recompress until the approximation accuracy determined by the accuracy calculation means reaches a predetermined value; It is characterized by being formed with the upper right side facing upward.

(Function) According to the present invention, the compressed digital signal qa4! is obtained by efficiently compressing the original digital signal group by operating the wood light equipment according to the method of the present invention. :Obtainable.

That is, if the compression means
A compressed digital signal is created by compressing based on a plurality of compression sections whose lengths in the measurement direction are unequal according to the degree of change in the value in the H.111 value direction. After that, the approximate viscosity calculation means calculates the approximation accuracy of the obtained compressed digital signal group with respect to the original digital signal station group, and if this approximation accuracy is not a predetermined value, the compression section addition instruction means determines that the approximation accuracy is a predetermined value. The compression means adds a new compression section to the compression section with low approximation accuracy and recompresses the compression section until the approximation accuracy is low.

As a result, a compressed digital signal group with high compression efficiency that responds well to changes in the values of the original digital signal group can be obtained.

〔Example〕

Embodiments of the present invention will be described below with reference to FIGS. 1 to 21.

As a group of 1.1 original digital signals, Fig. 1 J3
As shown in FIG. 2, image data consisting of the gradation of image 1 is targeted, as in the conventional case.

FIG. 1 t Digital CD group of the present invention)
It shows the strategy.

The digital signal group compression device 11 of this embodiment compresses a group of original digital signal stations whose values change with respect to one measurement direction using a plurality of compression sections having unequal lengths in the measurement direction. Compression means 12 for compression, and approximate accuracy calculation means 13 for calculating the approximation accuracy of the compressed digital 1+"i r intoxication obtained by the compression means 12 to the original digital signal.
and a compression section that issues a command to the compression means to add a new compression section to the compression section with low approximation accuracy and recompress until the approximation accuracy obtained by this approximate viscosity oil 0 means reaches a predetermined value. additional instruction means 4. The compression means 12 compresses the degree of change in the value of the original digital signal group in the measurement direction, which changes in value with respect to one measurement direction, that is, the degree of change in the value in the measurement direction of the image 1 in the X direction of the image data consisting of m degree gradation values. A change degree calculation means 15 for calculating the degree of change, and a compression section length determining means for determining the length of the compression section in the measurement direction, that is, the quality in the X direction, according to the degree of change in the image data obtained by the change degree calculation means 15. 16, and compression interval quality determining means] 6, and a zero compression means 17 for converting the original digital signal into one compressed digitized zero according to a predetermined method using the compression interval determined by the compression interval quality determining means. Furthermore, a CPtJ18 is provided that causes these means to operate in conjunction with each other.

Next, the operation of this embodiment will be explained.

In this embodiment, as shown in FIG. 3, the entire range of image data in the scanning direction is divided into unequal compression sections with different section lengths, that is, the lengths in the X direction, and each compression section is used to create an original image. It converts each group of digital signals into a compressed digital tenet.

This operation will be explained with reference to the general flowchart 1- in FIG. 4. When the compression processing operation is started, first, as shown in step 5T11, the image data input device 5 performs a density detection scan in the X direction JrjJ of the image 1. Then, the image data is read, that is, the original digital signal is read. Next, as shown in step 5T12, a compression section is set by the change degree section m15d3 and the compression section length determining means 16. In addition, the degree of change calculation means 15 calculates the degree of change in the values of the original digital signal station group, and the compression interval length determination means 16
Determine the degree of compression depending on the degree of change.

Next, at Stella 1S block, 3, No. 15 L [reduced block 1
In step 7, a compressed digital signal group is created according to a predetermined method using the J: compression interval. Next, step 5l-14
In the compression means 12, the approximation accuracy derivation means 13
The approximation accuracy for the original digital signal station of the compressed digital signal obtained by 41 months is determined, and this approximation accuracy is compared with a predetermined value set in advance for each compression section.

If the approximation accuracy falls within the predetermined value, the compression process is terminated; however, if it does not fall within the predetermined value, the process proceeds to step 5T15. , a new compression section is added within the compression section with low approximation accuracy, and the process returns to step 5T13 to perform signal compression!@filling, and then step 5T
What is the judgment of approximation accuracy in 14? 1.

This step ST→5T13→S "14 processing (Y, 5) is repeated until the approximate vi degree of the compressed digital Shingetsu reaches a predetermined value (directly).

There are ← methods for the constant compression period of the compression section in step 5T12. In this case, it is better to shorten the length of the compression interval for parts where the degree of change in value is large, and increase the length of the compression interval for parts where the degree of change in value is gradual, for better responsiveness to changes in the value of the original digital signal group. A corresponding group of compressed digital signals can be served.

However, in order to concretely set the JT: line section in Step 5T12, it is necessary to determine the section points that are the start and end points of each compression section, which are indicated by black circles on the X-axis in Fig. be able to (j) properly.

Selection of this interval point is preferably carried out by the change degree calculating means 15 and the compression interval length determining means 16 according to the flowchart shown in FIG. 5, for example.

That is, in step 5T21, a plurality of pixel 4i1 locations, ie, the X coordinates of pixels, where the degree of change in value is severe in the image data of FIG. 3 are selected as interval point candidates. Next, in step S-22, the set of interval point candidates selected as described above is narrowed down to a predetermined number of interval point candidates. The area point candidates are narrowed down to those that are spaced apart by a predetermined minimum image number or more. Next, in step 5T24, a number of new section point candidates corresponding to the number reduced by the narrowing down in step 5T23 are added to obtain the predetermined number of section point candidates.

A more specific example of the flowchart 1 through FIG. 5 is the broach 1 foot shown in FIG. 6.

To further explain this FIG. 6, step 5131
A plurality of interval point candidates are selected in . This selection is to select a location where the degree of change in image data is large, and as a selection method, it is preferable to select pixels that satisfy the following conditions (1) and (2>, for example).

Now, let the gradation value of the pixel data shown in FIG.
If D = g(il1) '(i), then (1)D -D<0 (i-1) (i) (2) wax(ID 1. ID l
) >C(i-1) (il where C is a constant (default value)) Locations that both satisfy the following conditions.

Conditions (1) and (2) are such that the digital signal of the pixel portion whose However, the condition f't is that the product of the absolute values of each change company exceeds the degree of severity of the change amount, which is a predetermined number C or more.

Next, in step 5T32 a3, the step 5T
For each section point candidate selected in 31, the following) (wax
(ID(1-1) 1. ID l > f
i) The performance (number of pixels is assigned in descending order). In this case, if there are multiple interval point candidates with the same performance L & j and it is necessary to select from among them, for example 1, [pixel X of
It is said to choose in order from the one with the smallest coordinate value.

Next, in step 5T33, a predetermined number (for example, 30) of −1-ranked interval points are selected from the set of interval point candidates.

Next, in step 5T34, it is determined whether or not the interval between the 30 selected section point candidates falls within a predetermined minimum number of pixels as a neighborhood condition. If the section point candidate is located nearby, it is determined to be YFS and Stella 1S-3
5, and if there is no one in the vicinity, the determination is No and the pre-section selection operation ends.

Next, in step 5T35, one of the two interval point candidates that are considered to be neighboring points is deleted, and the other is left as an interval point candidate. In this case, assuming that the interval point candidates whose X coordinates are 1 and j are now in the vicinity of each other, D - D (i-1) (i) ' and " (j-1) " D
(j) and keep the larger value and delete the other.

Next, in step 5T33, it is determined whether or not the total number of section point candidates selected so far is the predetermined number (30 in the case of tree-falling) determined in step 5T33, and Y
In the case of ES, the pre-section selection operation ends, and in the case of No, the process proceeds to step ST3□.

In this step 5l-37, the predetermined number (30
), the number of new interval point candidates that are not in a neighboring relationship with the already selected interval point candidates are added to
In the same manner as T32, select and add in order from the top, and return to step 5T36.

The operations in steps ST and ST3□ are repeated until the determination in steps G 3 and 6 becomes YES, and the pre-section selection operation is completed.

By determining the interval points in this manner, a compressed interval between each interval point is determined, as shown in FIG. This compression section let is set so that jJ3 has a short section length in the high frequency component region, and has a long section length in the low frequency component region.

Thereafter, in step 5113 shown in FIG. 4, a compressed digital signal is obtained using each compression section, and as a rest, a group of compressed digital signals in one scanning direction is obtained.

FIG. 7 shows the operation after step 5T13 in FIG. 4. That is, after reading the original digital signal g group in step 5T41 and setting the initial compression section in step 5T42, the process proceeds to step 5143.
A compressed digital signal 8 is obtained as shown below. In step 5T43, an equation is created for determining an approximation function for obtaining a compressed digital signal for each of the plurality of compression sections obtained as described above. Next, in step 5T44, a solution to the equation is found for each compression section to create a group of compressed digital signals.

Next, in step 5T45, the approximation accuracy (S/
N ratio) is calculated for each compression section. Thereafter, in step 5T46, it is determined whether the calculated approximation accuracy is within a predetermined value. If No, the process proceeds to step 5T48, and if YES, step 5T46 is determined.
Proceed to 47.

In this step 5l-47, a new section point is added to the compression section with low approximation accuracy to add a compression section. Thereafter, the process returns to step 5T43 and step 5T44 to calculate the compressed digital signal density again including the added compression section, and then calculates the approximate viscosity of the compressed digital signal signal updated in step 5T45, and in step 5T46 In step 1, it is determined whether the approximation accuracy falls within a predetermined value.

The operations in steps ST and 5TST 4743, 45, and 5T46 are repeated until the determination in step 5T46 is positive.

Next, in step 5148, the data of the calculated approximation function is stored as a compressed digital signal in a permanent memory for later use. I do this and finish.

There are various methods for setting the additional compression section at step ST4□ in FIG. A compressed section is added according to the method.

To explain FIG. 8, in step ST, 1, candidate compression sections are ordered in order to select additional section points.

This ranking GJ will be explained with reference to FIG.

FIG. 9 shows the already determined interval points ξ and ξ.

2 ξ3. Original digital signal 0 in the range of ξ, ξ5
In this embodiment, in each compression section between each section point, the compressed digital signal group in J3 is arranged in descending order of the degree of approximation M to the original digital signal group. Rank them. The approximation accuracy in this compression interval is as follows: ξ is the X coordinate of the interval point ξ, L(t) J'3 and 1(t) are the original digital signal station group and the compressed digital signal station at the X coordinate (1). Let it be the r8\A value of the group.

From this equation, if the approximate viscosity of the compression section is low (, L, this means that the value of the equation is large.

Next, in step 5T52, the No. 1 compression section from the ranked candidate compression sections is selected as a compression section from which an additional section candidate is to be selected.

Next, in step 5T53, in order to determine the method of adding interval points, it is determined whether the error variance of the compressed interval selected in step 5T52 is 4 or more, and YES is determined.
In the case of No, the process proceeds to step 5T54, and in the case of No, the process proceeds to step 5T61.

This error variance is determined by the following equation.

However, d V (tl = L (t) −1(t)′J
I・I C−Σdam/(ξi+1−ξi+1)t−%.

shall be.

When the error variance is 4 or more, as shown in FIGS. 10 and 11, this indicates a state in which the degree of change in the difference between the original digital signal group and the compressed digital belief group is large in the measurement value direction. Furthermore, when the 7f4 difference variance is less than 4, as shown in FIGS. 12 and 13, the degree of change in k between the original i'' digital input group and the compressed digital input;
The value indicates a state in which the value is inward. In this embodiment, if the error fraction is large, such as 4 or more, in step S-"
54 to step 5l-60 to add interval points to the part where the error is large, and if the error variance is small (less than 4), add interval points to the center of the compressed interval from step 5l-61 onwards. I try to say it.

On the other hand, in step 5T54, additional interval point candidates are ranked for points within the candidate compression interval.

In this case, in this embodiment, FIGS. 14(a) and (b)
As shown in , the original digital signal JI′ at the xi mark (i) other than the already determined interval points at both ends of the candidate compression interval
The values of the i group and the compressed digital signal @ group are i) and 1(
:) and the absolute difference L(i3-.111)l are determined and used as additional interval point candidates in order of preference.

Next, in step 5T55, one of the candidates obtained in the above manner is selected as an additional section point candidate.

Next, in step 5T56, it is determined whether or not there are already determined interval points in the vicinity of the unit pixel 0 on both sides of the added interval point candidate, and in the case of NO, step ST5□ If YES, the process proceeds to step 5T56.

On the other hand, in step 5T57, as shown in FIG. 15(a), the additional interval point candidates are separated from the already determined interval points ξ and ξi+2 on both sides by more than 2 pixels from I+1. As shown in (b), let the additional interval point candidate be an additional interval point ξ,,2,
Subsequently, the order of already determined section points with larger X coordinates is sequentially advanced.

In the other step 5T58, it is determined whether the already determined interval point near the additional interval point candidate is a single point, and if YES, step 5T59
If NO, the process proceeds to step 5T6o.

On the other hand, in step 5T59, a single point next to the candidate for an additional section point is made into a double multipoint point and is used as an additional section point. In this case, if there are single points t+-1- on both sides of the additional section point candidate, the smaller single point in terms of the X coordinate is used as the 2i112 multipoint point to be added as the additional section point. At J3 in Fig. 16(a), there is a single point, the already determined interval point ξ, only on the left side of the additional interval point candidate, so as shown in Fig. 1+1(b), An additional interval point ξi+2 is added using the interval point ξ· as a double multi-1+1 point. In the figure, multiple points are indicated by white circles. In addition, in FIG. 17(a), there are already determined interval points ξ5 which are single points on both sides of the additional interval point candidate.
, ξ, exists, ++2 ++3 The interval point ξ1+ with small X coordinate value as shown in (b)
2 is set as a double multipoint point, and an additional interval point ξ1+3 is added.

In the other step 5l-6o, FIG. 18(a)
As shown in FIG. 3, since the already determined section point next to the additional section point candidate is multiple section point ξ1+1°ξi+2, the additional section point candidate is canceled as shown in FIG.

Then, in order to set the next additional interval point candidate, the interval point candidate whose additional interval point candidate has been canceled is set in step 5T.
In step ST54, it is determined whether it is the last one among the ranked ones, and if YES, the process returns to step ST5T52, and if no, the process returns to step ST5. Returning to , the same additional section point setting process as described above is performed for the additional section point candidate of the next rank.

If the determination in step 5T53 is No, the process proceeds to Stella 1S 61, and as shown in FIG. 19, the midpoint of the candidate compression interval is set as an additional interval point candidate.

Next, in step S162, similarly to step 5T56, it is determined whether or not there is an existing section point near the additional section point candidate, and step N is performed.

In the case of YES, the process proceeds to step 5T63, and in the case of YES, the process proceeds to step 5T64.

In one step 5T63, an additional section point candidate is set as an additional section point.

In the other step 5-r64, similarly to Stella 1S, 8, it is determined 711 whether or not the already determined interval point in the vicinity of the additional interval point candidate is single;
If No, the process returns to step 5T52 to select the next candidate compression section; if YES, the process returns to step 5T6.
Proceed to step 5.

In this step 5T65, similarly to steps ST and 9, the single point next to the )0 addition interval point candidate is set as a double multipoint and an additional interval point. In this case, if there are single points on both sides of the additional interval point candidate,
The single point with the smaller X coordinate value is used as a double multipoint point and an additional section point.

By not repeating this compression processing operation for all pixels in the y direction, image data for the entire image tX11 is created.

The compressed digital creed group obtained in this way is sent to the host computer 6, and the print signal is edited with the character data of the character 2 sent to the host computer 6 through the character data human power device 4. The image is subjected to creation processing and printed by the printer 7 on the paper 3 as a reproduced image l1G118.

The image quality of this printed reproduced image @1a is of high quality almost equivalent to that of the original image 1.

This is the compression means 1 of the digital signal compression unit 5A and 11.
2, the compressed digital signal group created by the signal compressing means 17 is smoothed in the X direction based on the compressed intervals having non-uniform interval lengths determined by the change degree unit 9 means 15 and the compressed interval length determining means 16. A reproduced compressed digital signal is created by connecting the compressed digital signals so that the signal changes as shown in FIG. This is because the reproduced compressed digital signal S'3 is created by inputting the approximate viscosity to a predetermined value.

'J'eh straw, this re't-J, F reduction digital i d
This is because the signal changes almost the same as the original digital signal station consisting of image data.
Compared to the original digital communication station group, the reduced digital 4; The burden placed on the printer 6 and printer 7 is also reduced, which allows high-speed data processing to be performed whenever possible.

Next, each compressed digital signal 'i IJ J'; and each compressed digital CD, which are obtained by compressing the same group of l+7 digital signals 5] by the method of this embodiment and the conventional method, are compressed in the X direction. Figures 20 and 2 show the regenerated compressed digital Shingan rocks that are connected in a smoothly changing shape.
Compare using Figure 1. FIG. 20 shows the results obtained by the method of this embodiment, and FIG. 21 shows the results obtained by the conventional method. In both figures, the image data to be compressed is shown by the solid line.
It is composed of 256 pixels (one scale has two pixels), and the density gradation of image 1 is expressed in a digital scale for each pixel in the entire scanning direction range.

Therefore, in the method of this embodiment, as shown in FIG. The line section length is divided unevenly into sections indicated by black circles or white circles, and the compressed section is used to create a compressed digital signal. Yes, and if the variation in gradation value is very large,
It is also possible that the interval goodness is II OIT. , (In the old case where the section length was 0'', it is assumed that there are overlapped section points on the same -X coordinate (measurement direction).) In this example, as shown in FIG. There are 9 overlapping interval points (points indicated by white circles), which are divided into 49 compressed intervals in total.The broken lines in Fig. 20 indicate these points. It shows a group of reproduced compressed digital signals that connect the reproduced compressed digital tenets of each compressed digital signal (5, Q).In addition, in FIG. 20, the reproduced compressed digital signal ( (dashed line) overlaps, the image data that is the original digital belief group is given priority in the illustration.

In addition, in the conventional method, as shown in FIG. 21, the entire scanning direction range is equally divided into 181 compression sections (section points are indicated by black circles on xIl!lI) with uniform section lengths. A compressed digital C8 group is created. The broken lines in FIG. 21 indicate a group of reproduced compressed digital signals connecting the reproduced compressed digital beliefs obtained by re/1ing these compressed digital signals S3.

As can be seen by comparing FIG. 20 and FIG. 21, according to the method of this embodiment, the number of compression sections of the conventional method is reduced by about 1.
While compressing the 'C original digital signal 0 having 72 compression intervals, the obtained re-condensed digital signal group is as good as that of the conventional example having a large number of compression intervals, In response to changes in the values of the original digital signal, the bamboo bark responds with a roar. Therefore, according to this embodiment, the original digital signal group can be compressed extremely efficiently.

In the above embodiment, the compression target was the original digital signal group consisting of the gray level of each pixel of image 1g+1.
Any digital signal group whose value changes in one measurement direction can be compressed.

In addition, as a method for determining the length of the compression section in the measurement direction,
Methods other than those in each of the above embodiments may be adopted.

For example, if the intervals between the predetermined number of section point candidates determined up to step 5T24 in FIG. 6 are greater than or equal to the maximum number of pixels of the front paw, the intervals are equally divided within the maximum number of pixels. Alternatively, interval point candidates may be added and the compression interval may be determined using all these interval point candidates as interval points.

Alternatively, interval points may be added due to monotonicity exceeding a predetermined value. That is, in a monotonous section in which the gradation value only increases or decreases as the X coordinate increases, the starting point and The end point may be added as a section point candidate, and the compression section may be determined using all these section point candidates as section points.

The method and apparatus of the present invention are not limited to the above-mentioned embodiments, but can be modified in various ways as necessary.

〔Effect of the invention〕

Since the digital signal reduction method and apparatus of the present invention are constructed and operate in this way, it is possible to reduce the gap digital signal 2''i correction whose value changes in one measurement direction with good responsiveness to the change. It is difficult to compress compressed digital signals that change from time to time, but it can be efficiently compressed using a small number of compression sections, and subsequent processes such as Shinro Process 1III can be performed simply but at high speed. Qin Zhu can do such effects.

[Brief explanation of the drawing]

1 to 20 show an embodiment of the present invention, the digital signal ΣJ group compression method and the bag opening, FIG. 1 is a block diagram schematically showing the device of the present invention, and FIG. 2 is a block diagram schematically showing the device of the present invention. FIG. 3 is a block diagram showing the case where the method is used in a printer that performs image processing and image reproduction. FIG. Figures 4 to 8 show the cases in which
(Fig. 9 is a flowchart showing the autumn state of compression processing according to the method of the present invention, Fig. 9 is a similar diagram to Fig. 3 when selecting a candidate compression section to which an interval point is added, and Figs. 10 and 11 are 3 and 13 respectively show a case where the error variance is large, and FIG. 12a'3 and FIG.
) and (b) are explanatory diagrams in the case of finding additional interval point candidates at arbitrary places d, respectively, Figs. 15 (a) (b), Figs. (b) and Figure 18 (
a) and (b) are schematic diagrams explaining the method of setting additional interval points, FIG. 19 is an explanatory diagram for finding additional interval point candidates at the midpoint of the candidate compression interval, and FIG. 20 is the method of the present invention. 21 is a diagram similar to FIG. 20 showing a conventional example for comparison with the present invention,
FIG. 22 J3 and FIG. 23 are similar to FIG. 2 and FIG. 3, respectively, showing conventional examples. DESCRIPTION OF SYMBOLS 11... Digital signal g group compression device, 12... Compression means, 13... Approximate precision calculation means, 14... Compression section addition instruction means, 15... Change degree calculation means, 16.
... Compression section length determining means, 17... Signal compression means,
18...cpu.

Claims (1)

[Claims] 1) A digital signal group compression method in which an original digital signal group whose value changes with respect to one measurement direction is compressed using compression sections divided into a plurality of compression sections in the measurement direction, A compressed digital signal group is obtained by compressing the original digital signal group using a plurality of compression sections having unequal lengths in the measurement direction, and then the approximation accuracy of the compressed digital signal group to the original digital signal group is a predetermined value. A digital signal group compression method characterized by adding a new compression section within a compression section with low approximation accuracy until . 2) compression means for compressing a group of original digital signals whose values change with respect to one measurement direction using a plurality of compression sections having unequal lengths in the measurement direction; approximation accuracy calculation means for calculating the approximation accuracy of the compressed digital signal group to the original digital signal group; A digital signal group compression device comprising compression section addition instruction means for issuing a command to add a new compression section to the compression section and recompress it.