JPS5827146A - Interpolator for scanning line in photoengraving system for video image - Google Patents

Abstract

PURPOSE:To select an adequate interpolation method according to an image and to permit satisfactory photoengraving of a video image by making selection of any cubic convolution method possible, a bilinear method and a nearest neighbor method according to the contents of the images. CONSTITUTION:Means of interpolating scanning lines by a cubic convolution method, a bilinear method of a nearest neighbor method are provided and the calculations for interpolation are performed with logic circuits according to the quality of respective images, whereby the effective interpolation method can be selected freely. Any of ROMs 16a-16d is selected by switching an address selecting circuit 14 with a switch 13 for selecting interpolation methods. The ROMs 16a-16d are segmented to the kinds of the interpolation methods, the positions of interpolation ranges and the interpolation positions within the respective interpolation ranges, and are further discriminated to four regions including ''through''.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to scan line processing techniques in video prepress systems for producing film or print separations.

When comparing an image obtained from a television signal with a photograph taken directly with a camera, the quality of the former is significantly inferior to that of the latter. There are various reasons for this, but the unavoidable cause lies in the television system itself. In other words, in the standard television system, seven screens are composed of 325 scanning lines (the effective number of λ lines), so the screen size is A! If it is about R version, the scanning line density is 3. ,
There are two scanning lines and the scanning lines are visible. In order to solve this problem, conventional methods include wobbling, in which the electron beam is slightly waved by superimposing a minute alternating current on the deflection circuit of the cathode ray tube, and methods such as slightly shifting the focus of the electron beam to make the bright spot larger. However, all of them result in a decrease in resolution and degrade image quality. In contrast, the scanning line density in general printing ranges from 6 lines per page. Since it is a 2g book, there are no visible scanning lines, and the image quality is much better than prints made from television quality. A more effective way to deal with this problem is the interpolation method, which focuses on the correlation between scanning lines, and uses interpolation to form a new scanning line in the middle of the scanning lines, increasing the number of scanning lines. ing. Conventionally, as an example of incorporating an interpolation section into a plate-making device, JP-A-51!
, - No. 81117 discloses a television image plate making apparatus. However, the interpolation method disclosed in this device is a simple average value interpolation method and cannot be considered effective for all image information, and the number of interpolations is only interpolation between pixels/points. However, it can be said that the resolution is insufficient. In addition to the average interpolation method, there are also known interpolation methods that use a larger amount of data, such as the cubic convolution method. bilinear method.

The nearest keeper method is known. Among these, the cubic convolution method is the most advanced;
Therefore, we believe that if we use this cubic convolution method to interpolate scanning lines in TV image plate making, we will always be able to perform the best interpolation. It will be done. However, when interpolating a VC television screen, it was found that depending on the image structure, better results could be obtained by using the bilinear method or the nearest keeper method than by the cubic convolution method.

For example, in the case of an image composed of vertical and horizontal lines, better interpolation can be achieved by interpolating using the bilinear method. Therefore, an interpolation device that can employ only one type of interpolation method has the disadvantage that it cannot perform good interpolation for various images.

In addition, in the case of interpolation calculations that are generally performed using the Cupic convolution method, it is necessary to perform complex calculations taking into account the data of the second pixel, so the processing is usually done using computer software, but in this case, the Although it is simpler, it has the disadvantage of requiring a long processing time. Traditionally, video plate making equipment has been required to perform fairly high-speed processing, so when processing with software as described above, the data for each pixel that has been calculated is stored one after another in a buffer memory, and the calculations for seven screens are completed. When the calculation is completed, it is necessary to output the calculation results sequentially while synchronizing with the output device. However, the memory required for one screen has a large capacity of about 1,000 S megabytes or more, and the need for such a memory is extremely disadvantageous.

An object of the present invention is to solve the above-mentioned problems and provide a scanning line interpolation device that can select an appropriate interpolation method among the three methods of cubic convolution method, bilinear method, and nearest neighbor method depending on the image. It is.

The present invention relates to a video image plate-making apparatus for creating a printing film or printing separation plate from a television image, and a cupic convolution method.

A means of interpolating scanning lines by interpolation using the bilinear method and nearest neighbor method, and a cubic convolution method depending on the image content.

It is an enemy that provides a means for selecting either the bilinear method or the nearest neighbor method.

Still another object of the present invention is to provide a scanning line interpolation device for a video image prepress system that is quick and has good resolution by using a logic circuit constructed of elements suitable for high-speed processing.

The interpolation device of the video image plate making system of the present invention that achieves the above object is equipped with a logic circuit configured with elements suitable for high-speed operation in order to perform interpolation calculations at high speed, and is capable of inputting interpolation data in real time. The present invention is characterized in that it is provided with means for outputting.

Hereinafter, the present invention will be described in detail with reference to the drawings.

Here, first, an outline of an image interpolation method will be explained, and then embodiments of the present invention will be explained with reference to the drawings. Interpolation methods conventionally used in image processing include the nearest neighbor method and the bilinear method. The nearest neighbor method assumes that the data between adjacent scan lines does not change, and uses the data of the previous scan line as is during interpolation, while the bilinear method uses the data that changes linearly between adjacent scan lines. It shall be treated as such.

Therefore, the nearest neighbor method is effective for images consisting only of black and white converted into λ values, and the bilinear method is effective for images with halftones that change linearly, but many images Many of them have both large areas and mid-tone areas, and neither of these methods yields satisfactory results. Of course, the correlation between a point to be interpolated and surrounding pixels is greatest for the adjacent pixels, but furthermore, the correlation with pixels on the outer periphery cannot be ignored. On the other hand, since the nearest-neighbor method and the bilinear method ignore this correlation with pixels on the outer periphery, they are valid only under the above-mentioned special conditions. Various interpolation methods have been proposed that take into account the correlation with peripheral pixels, but the method adopted in the present invention is called the cubic convolution method, which takes into account the influence of peripheral /A pixels. These /6 pixel data are multiplied by a pixel coefficient representing correlation to find the sum, and the data at the interpolation point to be found is obtained. In the embodiment of the present invention, the standard television system, interlaced scanning, is switched to a progressive scanning system, and three IC scanning lines are interpolated between each scanning line to obtain one effective/92' scanning line. When exposing the film to light, the scanning lines are erased so that they overlap each other. At this time, the running linear density is about 73 lines/print for A5 size, which is about the same as that of ordinary printed matter, and can be put to practical use. Further, by performing the above-mentioned interpolation, the number of scanning lines is not only increased, but also the original data can be restored, so that the resolution increases from about 72 times to 3 times. Interpolation calculations are also required for conversion in the scanning direction. When shooting a TV screen, it may be necessary to turn the horizontal screen vertically, and each screen must be arranged so that the same screen is arranged horizontally and vertically on 7 screens. There may be times when shooting at different times. At this time, from the horizontally scanned screen, the pixel data necessary to form a scanning line with a density that can be used for vertical printing is determined by interpolation calculation, and this scanning line data is used to calculate the horizontal direction. By scanning, you can turn a horizontal image into a vertical image. Furthermore, when the image projection device for the Video 1III image used as the output expression of this system uses digital values for the king scan, and when the input signal from the prepress scanner uses digital values, all pixel data is obtained by interpolation calculation. There is a need to Each of the above three types of internal pressing methods has its special features, and the nearest impossible bar method is effective for monochrome images converted into Ω values, while the bilinear and modulus methods are effective for linearly changing halftone images. So, you can choose any method along with cubic compo I) kneeling method 5V
It has a C configuration.

Conventionally, there has been no idea of switching between these methods depending on the image quality, but according to the present invention, the most effective interpolation method can be freely selected depending on the image quality. In addition, in the present invention, the interpolation calculation up to t is performed by a logic circuit 118VC composed of reduction elements capable of high M calculations, making it possible to calculate in real time and eliminating the need for a buffer memory, making it possible to significantly simplify the configuration. can.

Next, details of embodiments of the present invention will be explained with reference to the drawings. In addition, in the video plate making system, the standard television method is separated into three primary colors R1G and BK.
, B, three systems of interpolation calculation circuits are prepared as described below.

FIG. 7 is an explanatory diagram showing the relationship between pixels of the original image and interpolated values. The number of pixels in the scanning line direction is sampled with a 4tfsc pulse and is set to 7 pixels and 1 pixel. Here f8° is the color subcarrier, 3. ! ;79. It is outside H2. According to the sampling theorem, the original image can be restored by sampling at a frequency that is one or more times the maximum frequency included in the image. In this example, the highest image frequency of the NTSC video signal is 6MH2 or less in general imaging equipment, whereas the fsc is /4'
Since the frequency is higher than MHz, the above conditions are fully satisfied and the restorability is good. In addition, in the case of signals with high resolution such as RGB format, the sampling frequency of f8c may be insufficient, so for example, A f8o (2/, lIg
MH2) If VC is used, the highest image frequency will be 10 MH2 or higher, allowing for more detailed digitization and interpolation. Further, in the case of 7 bling, which is a pulse of an even multiple of Vcf8°, each pixel is arranged linearly in both the iCx and y directions as shown in FIG.

FIG. 2 is a block diagram showing a circuit for A/D conversion of scanning lines and scanning method conversion. In the A/D converter/VC in Fig. 1, after sampling and digitizing the pixels of the scanning line as described above, the data is stored in the frame memory 2 by the write control circuit, 3Vc, and the read control circuit. Accordingly, the scanning lines of odd and even fields are read out alternately. In this way, it is possible to switch from the interlaced scanning method to the progressive scanning method, and to prevent bearings in which the spacing of the scanning lines is not uniform in some parts, resulting in a striped pattern.

The data at the point to be interpolated using the VcN of the present invention is the sum of the products of the adjacent /6 pixel data and the pixel coefficient of each pixel, as described above. This pixel coefficient represents the degree of correlation between the point to be interpolated and each pixel. If the axis and y-axis are taken in the main scanning direction and the sub-scanning direction, and the pixel interval in the can.

α, (x,y)=(1-21x12+lxM(1-21
y12+1ylJO≦lxl<1.0≦y<1
-=-f11α2(x, y) = (1-21X12+
lXl'+(4-81yl+51y12-1y 0≦l
xl<1.1≦l'yl<2 song-+21α5(
x, y) -(4-81Xl+51X12-IXI')
(1-21y12+ly 1≦lxl<2, o
5 lyl < 1 song... (5) α4 (x, y)
-(4-a1xl+5lx12-lxl')(4-al
yl+51y12-1y white 1≦lxl<2. t≦lyl
< 2 (4) α5 (x, y) = 0
2≦lxl, 2≦lyl (5
) Each pixel coefficient α1. ~α14 is calculated using the above formula, and its value differs depending on the location even if the interpolation point is inside the hatched area. The above explanation is an example using the cubic convolution method, but when applying the bilinear method or the nearest impossible bar method, by appropriately selecting α11 to α44, processing can be performed using the same method using the circuit in the cubic convolution method. can do. If each pixel data and pixel coefficient are expressed by Pij and α1j, the data at the interpolation point can be expressed by the following equation.

Generally, the sum of 76 pixel data is calculated using the above equation (6), but the interpolation point is xio 'x2 in Figure 1.
When it is on a pixel column such as 0' x50, the pixel coefficients of pixels on other columns are all 0, so the calculation becomes /. In this embodiment, in order to be compatible with an output device that receives all pixels as digital values as described above, a circuit that can perform the above-mentioned /6-pixel interpolation is implemented. The VC is configured so that interpolation and interpolation on pixel columns are performed by exchanging ROMs.

Further, in this embodiment, the space between adjacent scanning lines is divided into four equal parts, and nine scanning lines are obtained by interpolation. Seven of the original scan lines are on the original scan line, but since they were obtained by interpolation calculation, the processing is the same as for other scan lines. In this example, the scanning lines of the book are interpolated, but if there is no need for the book, it may be made two or three, and if there is a shortage, it may be three or six. In either case, /6 pixel coefficients are calculated for each interpolation point using equations (1) to (5), stored in the ROM, and used repeatedly for interpolation calculations. The pixel data read from the frame memory 2 is sent to the interpolation calculation circuit and compared to the line RAM VC, and is combined with the ROM in which the pixel coefficients have been written to obtain a required interpolation value. First, RA the t scanning lines at the top of the screen.
M K writing, X in Figure 1 for the leftmost 76 pixels. 0
Then, in the same manner, each pixel on the original scanning line is found. Next, the same operation is repeated to obtain each pixel in the fourth row, and then in the third row. Each pixel in the Dth row is determined to form a new scanning line of the book.

Furthermore, the /th row is erased and the jth row RAM K# is included, and nine new scanning lines including the third row are similarly formed in KL. In the same manner, the interpolation is completed until the bottom of the screen is reached.

FIG. 3 is an explanatory diagram showing a method of combining logic and ROM in one embodiment of the present invention. There are two methods of combining (1) and ROM: a method that uses seven types of ROM, and a method that does not change the RAM. In Fig. 3(A), K to find the interpolation point within the interpolation range As shown in Figure 3 (B) Vc, K ROM
−/ ~Stored in ROM −<7, RAM −'/ ToR
OM-/, RAM-, 2-bit ROM-2, RA
M-57 and ROM-,7, RAM-Q and ROM
This is possible by multiplying the corresponding pixel data and coefficients of -+ and calculating the sum. Next, to obtain the interpolated data in
-f's corresponding pixel data and pixel coefficients may be multiplied together to obtain the total sum. In this case, there are λ methods for combining RAM and ROM, and these are shown in Figure 3 (C1
The ROM is a ROM −/−ROM that stores pixel coefficients A' to D', respectively, as shown in FIG. 3(B)K.
- Only one set of pixels is used, and for correct -/- to -y, data is always written to l1II, which has a small scanning line number, and for the other seven pixels, as shown in ■ in Figure 3 (C). Change the order of coefficients A' to D' one by one and write ROM −/~ROM −
Set up 7 sets of things stored in Q, and once written to RAM
The pixel data of is left as is, and the third line is written into the RAM that had stored the first line. Below, X, , X4. The same is true for the interpolation point in X5, but the combinations I and ■ in the interpolation in The arrangement is the same as that of the Xl ROM. In other words, the same state is repeated every row. Figure 3 (According to C'l's combination, only 7 sets of ROMs are required, but the content of the theory must be rewritten every time the interpolation point is moved between adjacent scanning lines, and 17 g A bit selector is required.When rewriting the contents of the gland, the rewriting time becomes longer and the processing cannot be completed in time.

The combination (■) in the figure requires two sets of ROMs, but if you use a ROM element that is suitable for the total ROM capacity 11 required, it is sufficient to specify the address, so another ROM is required for that purpose.
There is no need to use M-separator 4, the configuration is simpler, and [RA
Since there is no need to rewrite the contents of M, processing time is shortened.

Next, the necessity of high-speed processing and the countermeasures taken in the present invention will be explained. As mentioned above, even if measures are taken to improve the image quality, printed matter made from television signals is not as good as those made from photographs taken directly with a camera, so it is intended for use in situations where many small-sized items are handled. Therefore, high-speed processing is required. Furthermore, in order to match the operating speed of a color scanner which is an output device of a video plate making device, the interpolation calculation speed requires a considerable processing speed per image. Furthermore, a video photographing device, which is the same output device, requires an even higher processing speed in view of the stability of the Fe3 blue deflection circuit. Among the above-mentioned requirements, the requirement for stability of the deflection circuit is the most severe. /To perform pixel interpolation calculation, calculate α1j-Pij by 76
It is necessary to calculate this sum by going around, and the interpolation points for 7 screens will be 77.2 points. Since it is completely unexpected that such a large number of calculations can be processed in a short time using a computer software program, the present invention uses a logic circuit composed of elements suitable for high-speed processing. Additionally, as a result, it became possible to send pixel data directly to the output device, which eliminated the need for memory for seven screens, significantly simplifying the configuration. The calculation of interpolated data also takes time to read from the memory, but the access time is 7 times for normal uoSRAM! rO ~ 200 n5ec, is required, so 2. for 713 calls. t~3. λ
μS, which does not satisfy the required conditions. Therefore, in the present invention, the same calculation circuit is divided into q circuits and parallel calculation is performed in an attempt to reduce this time.

FIG. 7 is a diagram showing the configuration of the interpolator circuit.

In the figure, for the reason mentioned above, the input buffer S, the line logic 1 multiplier, and the adder D are each provided with a digital circuit. Pixel data read out from the frame memory in a sequential scanning manner is stored in input buffers 3a to 3a.
The data is written to the line RAMs 7a to 7d via Sd. This writing control is performed by AV, BW, CW, DW.
This is done by the write control signal. The read control is
The address signal passes through the address buffer B to each RAM K.
This is done by adding lines RAM7a to 7.
Is the data in d a multiplier and an adder? Sent to a=qd.

On the other hand, the data in the predetermined 117') ROM passes through the pixel coefficient buffer g to the multiplier and adder? The calculation is performed by adding a=9rlVc. The combination of the line number of the scanning line and the line number of the ROM is performed by the combination (C1) in FIG. and output via path buffer //.

FIG. S [A] is an explanatory diagram showing the progress of interpolation is in the upper left corner of the screen. In the first step VC, multipliers and adders 9a, 9b, 9c, ? dV
C is "11, PI3" from FLAM 7a ~ 7d
15 ” 14 is also α11.α12 from the ROM.
．． α15. α14 is sent in and multiplication is performed respectively. In the third step, RAM 7
From a to 7d, P21. P22゜P25. P24
But from the ROM it is α21. α22. α25. α24 is sent to a multiplier and an adder to perform multiplication and addition with previously calculated data. Multiplication and addition are performed in the same manner in the third step, and PA, PB, and PCoPD are obtained in the third step. Next, when these data are sent to the adder/θ and the sum is determined, the resulting value becomes the interpolated data.

Figures 6 and 5 show R at each interpolation method and each interpolation point.
FIG. 2 is an explanatory diagram showing the configuration of OM and a reading method thereof. R.O.
The value of M differs depending on various interpolation methods and the position jtK inside the diagonal line shown in FIG. 7, and the interpolation range x1. x2. x5. The combination of each row is different. Once the interpolation position is determined, the above-mentioned (1) to (
The pixel coefficients can be determined using equation 5), and a ROM can be manufactured by writing this data as shown in FIG. In the interpolation calculation, once the interpolation method is selected, the necessary ROM is read out according to a predetermined sequence and the calculation proceeds automatically.') 1-No error. The types of interpolation methods are the cubic convolution method, bilinear method, and nearest neighbor method mentioned above, but there are 11 methods including the through method without interpolation.
The interpolation method is selected by the interpolation selection switch /3 shown in Figure 6, and then the address selection circuit /G selects the ROM/
A predetermined position from Aa to /Ad is selected. The configuration of each of ROM#a to /Ad is shown in FIG. 6, and in this case, the type of interpolation method described in E, the position of the interpolation range txl, x2. x5. x4. and the interpolation position 'xDO' xlo 1 X20 ,X in each interpolation range
It is divided into 50 areas, and each area is further divided into 5 areas, A/, B/, c/, and D in Fig. 3(B).
The pixel coefficients of any (I) row are stored. Inventively, the interpolation position is changed by replacing the ROM.For example, X on the original pixel column. g, X
The interpolation points of 10 *

In order to transfer to X52K, it is sufficient to replace the ROM in which the pixel coefficients determined by tt according to equation (6) are recorded. In order to increase the number of interpolation columns, it is necessary to add a ROM having a configuration similar to that shown in FIG. 6 and recording the pixel coefficients of each column. Also, the interpolation property, that is, the scanning l\llj! caused by interpolation! To increase or decrease the number, it is necessary to change the configuration of the ROM (7') in FIG.

To give an example of the data written in the ROM in this embodiment, the interpolation range of xl includes /Aa and α11. α2. .

α51. α41;/OtbVrα12. α22. α52.
α42;/6cVcα13・α25・α older brother・α45;/
AdVC α14, α24, α54, α44 are written, and the interpolation range of X2 is α14, α24, α in /Aa.
54・α44;/AbVCα11・α21°α51°α
41;//, α12. to c. α22Iα52. α42;/
6d α15'α25-α55. α43 is written.

FIG. 5 is an explanatory diagram showing the interpolation order of each interpolation point constituting seven screens. The interpolation calculation at each interpolation point is performed by multiplying the pixel series number of each pixel by the pixel data and calculating the sum, as explained using FIG. As shown in Figure 7, the interpolation is performed at X1 between the second and third rows at the top of the screen.
Interpolation position X at the top of the range. 7 scanning lines X to find the pixel data of the interpolation point from left to right. complete. The number of pixels in the original scanning line is 79, but since the number of interpolation points is 7 and 3, 31 pixels (0,114) are lost.

Next, the ROM is switched and Xl is found in the same manner, and X2 and X5 are found in the same manner. After that, the interpolation points of XO* It also changes. Thereafter, in the same manner, the interpolation calculation for the entire screen ends at X481Vc. The number of initial scanning lines is 4, and the interpolation position is 1, so 3 lines (θ, 62) are lost at the top and bottom.

Details [As explained above, in the present invention, the most suitable interpolation method can be selected depending on the content of the image, so the quality of the projected image can be significantly improved, and the printed separation created from this can be It was possible to significantly improve the image quality of printed matter printed using the plate. Interlaced scanning can be converted to progressive scanning to avoid bearings, and interpolation can increase the number of scan lines to make them invisible and improve resolution. In addition, high-speed processing M
By constructing an extremely high-speed interpolation calculation circuit using a logic circuit composed of elements suitable for A, it was possible to construct a stable video prepress system that is highly operable and practical. Furthermore, the present invention makes it possible to eliminate the need for large-capacity memories and many selectors, and has a great effect in simplifying the circuit configuration.

Furthermore, the present invention is not limited to the above-mentioned embodiments,
Many variations and changes are possible. For example, the present invention 1'c
In this case, three types of interpolation methods can be selected, but if it is not necessary, two or seven types can be selected.Although the number of scanning lines was multiplied by seven by interpolation, other multiples can be used. It is also easy. Furthermore, although the purpose of the present invention is to provide interpolation #2tf suitable for video plate making equipment, similar effects can be expected by utilizing the present invention to improve image quality in other applications. can.

[Brief explanation of the drawing]

Figure 1 shows the cubic convolution method [7
An explanatory diagram showing the relationship between a set of pixels and an interpolation point, Figure 1 is a circuit diagram of scanning line A/Df conversion and scanning method conversion, Figure 3 (All
An explanatory diagram showing the relationship between each pixel and the interpolation position. Figure 3 (B) shows 7 sets of pixel coefficients stored in ROM K. Figure 3 (C) shows the scanning line to be written in RAM Ic and the ROM at each interpolation position. A table comparing the present invention and other methods of writing pixel coefficients, Fig. G is a block diagram showing an interpolation calculation circuit, Fig. S is an explanatory diagram of an interpolation calculation method for seven pixels, Fig. 6 is a ROM surrounding circuit diagram,
FIG. 7 is an address configuration diagram of the ROM, and FIG. S is an explanatory diagram of the /side force interpolation order and interpolation range. λ...Frame memory, S...Input buffer, 6...Address buffer, Ri...Line RAM
. g...pixel coefficient buffer, de...multiplier and month/date calculator, 10...adder, ~~・V~layer kahyakuba bad l/''・
... bus buffer, /3 ... interpolation switch, /lI,
/3...Address selector, /6...Line ROM
0 ＼ N (I) Dimensions (1 risk 3 hours) γ■
al11beζ

Claims (1)

[Scope of Claims] 1. In a video image plate-making device that creates a printing film or a printing separation plate from a television image, interpolation is performed using a cubic convolution method, a pi-linear method, and a nearest non-/<- method M. A scanning line interpolation device is provided with a means for interpolating a scanning line and a means for selecting one of a cubic convolution method, a bilinear method, and a nearest neighbor method depending on the content of an image. 2. The scanning line interpolation device according to claim 1, wherein interlaced scanning is converted to progressive scanning before scanning line interpolation to prevent bearing. 3. A patent claim characterized in that the position of the interpolation point in the cubic convolution method is set on a pixel column between the original scanning lines, and the interpolation calculation is performed using only pixel data of the pixel column. The scanning line interpolation device according to item 1. 4. Means for storing each pixel coefficient for interpolation of cubic convolution method, pi-by-linear method, and nearest-Naher method, and means for selectively reading out pixel coefficients for a specified interpolation method from this storage means. and a RAM that stores pixel data of nine adjacent scanning lines.
2. The scanning line interpolation device according to claim 1, further comprising an interpolation circuit for performing interpolation calculation from the pixel coefficients and pixel data. 5. A video image plate-making device Vc that creates a printing film or a printed separation plate from a television image is equipped with a logic circuit composed of elements suitable for high-speed operation gK in order to perform interpolation calculations at high speed, and performs interpolation calculations in real time. A scanning line interpolation device characterized in that t\jm is provided with means for outputting interpolation data. 6. Sequentially combine rows of pixel coefficients consisting of 9 rows and columns
7 different combinations are made and stored in ROM K, and the pixel coefficients of any set in this ROM and the pixel data of the adjacent scanning line are stored. ,/Scanning line segment force interpolator, pixel data of a new scanning line is stored in the RAM from which the data has been erased after erasure is completed, and the configuration of the interpolation calculation circuit is simplified. 14°Z The interpolation calculation circuit stores the pixel coefficients of each interpolation method, the RAM that stores the pixel data of the scanning line, and the pixel coefficients and pixel data. 7. A scanning line interpolation table 1 according to claim 6, characterized in that a multiplier and an adder for performing interpolation calculations are provided.