JPH01142878A - Method and device for interpolating digital image - Google Patents

Abstract

PURPOSE:To obtain picture quality more excellent than an ordinary example by using an interpolation factor obtained by supplementing a 2nd factor determined by an interpolating position to a sinc function. CONSTITUTION:The interpolating position is set up as an approximate center, 2mX2n image data are selected so as to surround the interpolating position and distances (x), (y) up to the image data selected from the interpolating position are found out. Then, the values of prescribed factors omega(x), omega(y) uniquely determined from the found distances (x), (y) are found out. Preferably, the values of the prescribed factors omega(x), omega(y) are found out in accordance with the shown equations I. Then, interpolation factors I are found out in accordance with the equation II. Respective products of the interpolating factors I and image data corresponding to the interpolating factors are found out and the final interpolation image data are formed by the sum of obtained products 2mX2n.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to a digital image interpolation method and apparatus,
In particular, the present invention relates to a digital image interpolation method and apparatus for forming interpolated image data to be provided to an image output device.

[Prior art] Conventionally, as this type of interpolation method, nearest neighbor pixel approximation (n
earth neighbor method), two-dimensional linear approximation (
bilinear method), cubic B-spli
There are ne approximation, 5 inc interpolation method, etc. For details, see “
Digital Image Processing''
William K. Pratt (published by John Wiley & Fiveons, Inc., 1978) and others. However, in the above-mentioned conventional example, when the image was enlarged more than 4 to 5 times, the sharpness of the image deteriorated extremely. Furthermore, depending on the magnification, image quality deteriorated, such as dark and light stripes appearing at pixel intervals in the original image or a tiled image. In this respect, the 5ine interpolation method is theoretically considered to be optimal. However, the 5-inc interpolation method has a problem in practicality in that it is necessary to perform interpolation using all input pixels.

[Problems to be Solved and Attempted by the Invention] The present invention eliminates the above-mentioned drawbacks of the prior art, and its purpose is to provide a digital image interpolation method and apparatus that can put the 5-inch interpolation method into practical use. It is about providing.

[Means for Solving the Problems] In order to achieve the above object, the digital image interpolation method of the present invention includes the steps of storing a plurality of image data in integer coordinates of orthogonal coordinates X and Y; a step of setting an arbitrary interpolation position on the real number coordinates of X and Y, a step of selecting 2m×2n pieces of image data so as to surround the interpolation position with the set interpolation position as the approximate center, and the selection from the interpolation position. a step of calculating distances x and y to the image data,
The distance found above is 1! A predetermined coefficient ω(x) from Itx,y,
Step of calculating the value of ω(y) and the calculated distance 1affx
, y and the values of coefficients ω(x) and ω(y), the interpolation coefficient I
The process of obtaining πX πy according to the following formula, and the obtained 2
The outline thereof is to include a step of forming interpolated image data by the sum of m×2n products.

Preferably, the values of coefficients ω(x) and ω(y) are expressed as follows: One aspect of this is to obtain the following.

Preferably, one aspect thereof is that m=n=8.

Preferably, the interpolated image data is formed using a temporary interpolated image obtained only for the distance lax or y. ! One aspect of this is to perform calculations regarding Iy or X.

In order to achieve the above object, the digital image interpolation device of the present invention includes an image memory that stores a plurality of image data at integer addresses of orthogonal addresses an interpolation position setting means for setting an interpolation position; an image data selection means for selecting 2m x 2n pieces of image data surrounding the interpolation position with the set interpolation position as the approximate center; A distance calculating means for calculating distances x and y to the image data, and a predetermined coefficient ω from the calculated distances x and y.
(x), ω(y), and an interpolation coefficient ■ from the determined distances x, y and the coefficients ω(x), ω(y) according to the following equation, πX πy. and an interpolated image data forming means for forming interpolated image data by taking various types of the obtained interpolation coefficient (1) and image data corresponding to the interpolation coefficient, and forming interpolated image data by the sum of the obtained 2m×2n products. The outline is as follows.

[Operation] In such a configuration, for example, orthogonal coordinates X, Y
Set an arbitrary interpolation position on the real number coordinates of the orthogonal coordinates X and Y, and surround the interpolation position as the approximate center of the interpolation position. 2m×2n pieces of image data are selected as follows.

Preferably m=n==8. Next, the distance x+y from the interpolation position to the selected image data is determined.

Next, the values of predetermined coefficients ω(x) and ω(y) that are uniquely determined from the distances x and y determined above are determined. Preferably, the values of the predetermined coefficients ω(x) and ω(y) are determined according to the following equations.

Next, the distance found above is 1! lx, y and coefficients ω(x), ω
The interpolation coefficient ■ is calculated from the value of (y) according to the following equation, πX πy. In this way, predetermined coefficients ω(x) and ω(y) determined by the interpolation position are added to the 5inc function. Then, each block of the obtained interpolation coefficient (1) and the image data corresponding to the interpolation coefficient is taken, and the final interpolation image data is formed by the sum of the obtained 2m×2n products. Preferably, the interpolated image data is formed by inputting the temporary interpolated image data obtained in accordance with the above for only the distance X or y. It is determined by performing the above calculation for y or X.

[Description of Embodiments] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

[First Embodiment] FIG. 1 is a block diagram of a digital image interpolation device according to a first embodiment of the present invention. The first embodiment relates to one in which an NTSC video signal of a television is digitally converted, interpolation processing according to the present invention is performed, and the resulting interpolated image data is output to an image printer.

In FIG. 1, 1 is a video signal source, which generates an NTSC video signal VS used, for example, in television. A luminance signal extraction circuit 2 extracts a video luminance signal VLS from the video signal vS using, for example, a known color subcarrier removal filter. Reference numeral 3 denotes a synchronization signal extraction circuit, which extracts a video frame synchronization signal V from the video signal vS using a known circuit in, for example, a conventional television receiver.
FDS (vertical synchronization signal) and video scanning line synchronization signal V
Extract HDS (horizontal synchronization signal). 4 is a sampling period generating circuit, which generates the extracted synchronization signal VFDS.
, based on VHDS, for example, the video luminance signal VLS is converted to A/
A data sampling signal DSS (for example, 640 pixels/scanning line) for D conversion is generated. 5 is an A/D conversion circuit which A/D converts the input video luminance signal VLS and outputs digital luminance data LD. 6 is a γ conversion circuit which converts the input luminance data LD to an appropriate density level and outputs density data ND. In general, a television signal is a signal having a γ characteristic of (1/2.2), which matches the characteristics of a display surface fluorescent material for image reception.

Therefore, the γ conversion circuit 6 converts the luminance data LD having this characteristic.
is linearly converted into density data ND having an appropriate density level matching the printer characteristics. A frame memory 7 stores one page of input density data ND. A write address control circuit 11 controls the write address WAD and write timing of the density data ND to the frame memory 7. Specifically, the data sampling signal DSS is counted to calculate which pixel position (x coordinate) it corresponds to in the video scanning line, and the video scanning line synchronization signal VHDS is counted to calculate which pixel position (x coordinate) it corresponds to in the screen. It is calculated whether it corresponds to the th video scanning line (Y coordinate).

With the above configuration, each density data ND (image data GD
) are written in a continuous manner onto the integer addresses of the frame memory 7 in synchronization with the data sampling signal DSS.

FIG. 6 is a diagram showing one storage mode of image data in the frame memory 7. As shown in FIG. In the figure, a series of image data G
D (density data ND) is stored as raster data corresponding to a printing scanning line (raster). For example, first, the first pixel, second pixel, . . . , Nth pixel of the first rask are stored in order.

If necessary, a margin for the first rask is stored between the first rask and the second rask. This simplifies the hardware control of the X and Y addresses. Next, the first pixel, second pixel, ..., Nth pixel of the second rask are stored in order,
In this way, finally the first pixel, second pixel of the Mth raster, etc.
, and the Nth pixel are stored in order. In this case, the frame memory 7 continuously stores image data GD having a number of pixels corresponding to a power of 2, for example, for each rask. The size of one image stored in this way is, for example, 640 (pixels/raster)
x480 (raster/image), and the memory space for this is 1024 x 512 = 512 bytes.

In FIG. 1, reference numeral 10 denotes an image printer. When starting printing, it sends out a print page synchronization signal PPDS indicating the top of the page, and when printing, it sends out a series of print scanning line synchronization signals PSDS (sub-scanning synchronization signals ) and a print pixel synchronization signal PGDS (main scan synchronization signal) in each print scan line. These synchronization signals are input to an interpolation processing circuit 8, which will be described later, via a printer I/F circuit 9. 8 is an interpolation processing circuit which sets an interpolation position (real number coordinates) by counting various synchronization signals from the image printer 10 and also outputs synchronization signals (various carry signals) for updating the integer read address of the frame memory 7. CA
R), and together with these, the sum of the products of the interpolation coefficient I and the original image data (interpolated image data HGD) is generated.

A read address control circuit 12 updates the integer read address RAD of the frame memory 7 in synchronization with various carry signals CAR from the interpolation processing circuit 8. With the above configuration, the image data GD on the frame memory 7 is interpolated by the interpolated image data HGD and output to the image printer 10. Reference numeral 13 denotes a sequence control section, which controls the operating procedure of the entire apparatus.

FIG. 3 is a flowchart of the image data processing procedure by the sequence control circuit 13 of the first embodiment. In the figure, in step S1, it is determined whether or not there is a print instruction from an operation unit (not shown).

If there is no print instruction, the process advances to step S2 and other processing is performed. Other processing includes, for example, the image printer 1
This refers to checking the status of the video signal source 1 (such as power off) or checking the status of the video signal source 1 (whether the signal is being transmitted or not). If there is no abnormality in this check, the process immediately returns to step S1. Further, if there is a print instruction in step S1, the process advances to step S3, where the write address control circuit 11 is initialized, and then the density data ND write function is activated to store a series of density data ND in the frame memory. Start writing to 7. In step S4, it is determined whether writing to the frame memory 7 is complete. If writing is not completed, step S5
, and perform other processing in the same manner as described above. When the writing of all density data ND is completed, the process advances to step S6, and the data writing function of the write address control circuit 11 is disabled. In step S7, it is determined whether the image printer 10 is ready to start printing. If it is not ready to start printing, the process advances to step S8, and other processing is performed in the same manner as described above. Further, if the printing can be started, the process advances to step S9, and the interpolation processing circuit 8 is initialized.

Specifically, setting of each constant value (initial value, increment value) in the differential analysis circuit (DDA) 14 and DDA 15, which will be described later, clearing various buffer groups (F I FO, latch, etc.), and data flow (not shown) Initializes control counters and clears various gates, etc.

In step S10, the read address control circuit 12 is initialized and then placed in a data read operation enabled state.

Specifically, the counter and the like that hold the image data read address RAD of the frame memory 7 are initialized, and the first image data GD is made readable. step sit
Now, the printer I/F circuit 9 is initialized. in particular,
Reception of control commands and status signals with the image printer 10 initializes the I10 circuit for control information necessary for transfer, etc., and sends and receives the information, etc., as well as various synchronization signals related to image interpolation processing and transfer. initialization, and initialization of gate circuits and the like related to these signals are performed. In this way, from now on, the read address control circuit 12 and the interpolation processing circuit 8 operate in synchronization with the operation clock signal (various synchronization signals) from the image printer 10, and the image data GD is transferred from the frame memory 7.
are sequentially read out, subjected to interpolation processing, and the generated interpolated image data HGD is sent to the image printer 10 via the printer I/F circuit 9. On the other hand, the sequence control circuit 13 waits for printing completion in step S12. That is, it is determined in step S12 whether or not printing has been completed, and if printing has not been completed, the process advances to step S13 and other processing is performed in the same manner as described above. When printing is eventually completed, the process proceeds to step S14, where the image data read function of the read address control circuit 12 is disabled.

FIG. 2 is a block diagram showing details of the interpolation processing circuit 8 of the first embodiment. In the figure, 14 is D D A (
Digital Differential Anal
yzer) and performs digital differential analysis of the printer sub-scanning direction address (Y coordinate). Specifically, it generates a sub-scanning carry signal VCAR for updating the integer successive address (Y coordinate) of the frame memory 7, and also generates a sub-scanning carry signal VCAR for updating the integer read address (Y coordinate of the interpolation position).
coordinates). 15 is a DDA, which performs digital differential analysis of the printer main scanning direction address (x coordinate).

Specifically, the integer read address (x
It generates a main scanning carry signal HCAR for updating the coordinates), and also generates a real address (X coordinate of the interpolated position) located between the integer read addresses. Reference numeral 16 denotes a group of raw data scanning line buffers, which can store seven scanning lines of image data GD sequentially read out from the frame memory 7 in parallel. Reference numeral 17 denotes a sub-scanning direction interpolation table group, which is set to obtain the product of the temporary sub-scanning direction component interpolation coefficient 7 and the input image data for the eight image data GD arranged in one row in the sub-scanning direction. Each generates a predetermined value of a coefficient IY based on the distance y between the interpolated position and each image data, and calculates the product of the value of the coefficient IY and the image data.

Reference numeral 18 denotes a sub-scanning direction interpolation adder group, which calculates the sum of the multiproducts by the interpolation table group 17 and obtains the sum of the products of IY and input image data as a temporary interpolation result in the sub-scanning direction.

19 is a main scanning direction interpolation data buffer group, which stores the temporary interpolation result IY of the output of the sub scanning direction interpolation adder group 18;
Seven total products of the input image data and the input image data can be sequentially stored in the main scanning direction.

20 is a group of interpolation tables in the main scanning direction, in which interpolation positions are set in order to obtain the sum of the products of the final interpolation result I and the input image data for the 8 interpolation coefficients Iy arranged in a row in the main scanning direction. A predetermined value of coefficient IX is generated based on the distance x between each interpolated image data, and the value of coefficient 8 is multiplied by the above-mentioned temporary interpolation result. Reference numeral 21 denotes a group of interpolation adders in the main scanning direction, which calculates the sum of the multiproducts of the interpolation table group 20 to obtain final interpolated image data V. Find ut.

FIG. 13 is a diagram showing tables 17 and 20 of the embodiment. These are realized by a ROM, and the upper side of the address input corresponds to interpolation position information (ΔX or Δy), and the lower side of the address input corresponds to the value of the input pixel. The contents of the table are preset as the product of the interpolation coefficient at the interpolation position corresponding to each address and the corresponding pixel value.

Next, details of the operation of the interpolation processing circuit 8 will be explained.

When the page synchronization signal PPDS is input to the DDA 14, the selector 14-4 changes to a mode in which the contents of the initial value register 14-2 are selectively output. As a result, the adder 14-5 outputs the initial value of the initial value register 14-2 and the increment value register 14.
-1 increment value is added and the addition result is output. Next, when the first scanning line synchronization signal PSDS is input, the adder 14
The addition output of -5 is taken into the current value register 14-3, and at the same time the selector 14-4 changes to a mode for selectively outputting the contents of the current value register 14-3.

Therefore, from now on, each time the scanning line synchronization signal PSDS is input, the contents of the current value register 14-3 are updated by the value obtained by adding the increment value to the previous value. Also, each time an addition carry occurs in the adder 14-5 during this update cycle, a carry signal VCAR is generated and outputted to the read address control circuit 12. From the above, the output of the adder 14-5 can be considered to be the remaining fractional part when the carry is normalized to 1, excluding the carry signal VCAR. Therefore, when the distance of the sampling interval in the sub-scanning direction of the video signal VS is considered to be a unit distance @1, the output of the adder 14-5 is the real number distance 1 measured in the sub-scanning direction from each sampling position.
It represents y.

Similarly, when the scanning line synchronization signal psDS is input to the DDA 15, the selector 15-4 changes to a mode in which the contents of the initial value register 15-2 are selectively output. As a result, the adder 14-5 adds the initial value of the initial value register 15-2 and the increment value of the increment value register 15-1 and outputs the addition result.6 Next, when the first pixel synchronization signal PGDS is input, the addition The addition output of the unit 15-5 is taken into the current value register 15-3, and at the same time, the selector 15-4 changes to a mode for selectively outputting the contents of the current value register 15-3. Therefore, from now on, each time the pixel synchronization signal PGDS is input, the contents of the current value register 15-3 are updated by the value obtained by adding the increment value to the previous value. Also, when an addition carry occurs in the adder 15-5 during this update cycle, a carry signal HCAR is generated each time, and this is sent to the read address control circuit 1.
Output to 2. From the above, the output of the adder 15-5 can be considered to be the remaining fractional part when the carry is normalized to 1, excluding the carry signal HCAR. Therefore, when the distance of the sampling interval in the main scanning direction of the video signal vS is considered as a unit distance 1, the output of the adder 15-5 represents the real number distance X measured in the main scanning direction from each sampling position.

FIG. 5 is a schematic diagram illustrating the operation mode of the DDA of the first embodiment. FIG. 5 shows, as an example, the operation of the DDA when an original image is enlarged by 130% and subjected to interpolation processing. 1
When enlarging by 30%, a value of 1/1.3 is expressed in binary and 15 bits below the decimal point are used as the increment value. When the decimal part is expressed in hex, it is "313B". On the other hand, the initial value becomes the current value (decimal part) when the 8th carry signal CAR is generated by adding the increment value "313B".
Set a value such that the value is "0". In this case, the initial value is 13B2". Therefore, in FIG.
A relationship is shown in which the first carry signal CAR is generated when the first 10 clock signals are input from the 0 side. Here, if the clock signal on the image printer 10 side is the pixel synchronization signal PGDS, then FIG. 5 shows an example of the operation of the main scanning direction DDA 15. This shows a relationship in which eight pieces of input image data GD are read out during output. Furthermore, if the clock signal is the scanning line synchronization signal PSDS, then FIG. 5 shows an example of the operation of the sub-scanning direction DDA14. The relationship is shown in which eight lines of input image data GD are read out. Note that it is not necessary to perform the addition using the FULL 15 bits, and the addition may be performed using the upper several bits.

In FIG. 2, the read address control circuit 12 is reset by the page synchronization signal PPDS, and thereafter the DDA
The lower read address (main scanning direction address) of the frame memory 7 is obtained by counting the carry signal HCAR from 15.
and carry signal VCAR from DDA14.
is counted and the upper read address of the frame memory 7 (
(sub-scanning direction address). The read address signal RAD sent to the frame memory 7 is a combination of the upper and lower read addresses.

The raw data scanning line buffer group 16 includes seven shift registers (PIFOs 16-1 to 16-7), and the image data GD for seven scanning lines is read out from the frame memory 7.
Save up. That is, the image data GD read from the frame memory 7 is input to the FIFO 16-1 and shifted by 1024 pixels. Next, the output of FIFO16-1 is FIF
It becomes the input of O16-2, and in the same way, PIF016
The output of -6 becomes the PIFO16-7 (7) input. Furthermore, the image data GD input to PIFO16-7 is 1
When the image data is shifted by 024 pixels, the image data GD of the eighth scanning line is read out from the frame memory 7 in synchronization with this. If this 8th scanning line read raster is the current read raster, then any FIFO 16-i (i=1.2
．． =・.

The content stored in 7) is the read raster i lines before.

The sub-scanning direction interpolation table group 17 includes eight look-up tables (LUTs) 17-1 to 17-8. Each LUT inputs the image data GD of the corresponding read raster and the real distance y (output of the adder 14-5) measured in the sub-scanning direction from the current sampling position, and generates the density of the image data GD at the interpolation position. Power interpolated image data H
The product of the contribution rate (partial interpolation coefficient Iy in the sub-scanning direction) indicating how much contribution should be made from the sub-scanning direction to the density of GD and the corresponding input pixel value is output.

For example, the LtJT 17-1 uses the image data G Do of the current read raster and the current interpolation position y of the output of the adder 14-5 as address inputs.
17-1 is the value of a predetermined interpolation coefficient Iyo (partial interpolation coefficient) corresponding to the current interpolation position y and the corresponding image data GDO
The product value 1yoGDo with the value (density) of 1yoGDo is internally generated and output.

Similarly, LUT17-8 is set to 7 of PIFO16-7 output.
The image data G D before the line and the interpolation position y of the adder 14-5 are used as address inputs, and thereby the L
tlT17-8 is the value of the predetermined coefficient 13/7 (
A value iy, G[), which is the product of the partial interpolation coefficient) and the value of the image data GD7, is internally generated and output. In this way, the contents of the image data GD are sequentially updated in synchronization with the generation of the carry signal HCAR, and the above-mentioned partial interpolation coefficient Iy
The calculation of the product of n and input image data GD is performed sequentially. On the other hand, the contents of the read raster are sequentially updated one line at a time in synchronization with the generation of the carry signal VCAR. During this period, scan synchronization signals PSDS are normally generated at a rate of one or more (1, 2, 1, 3, etc.), and the interpolation position y in the sub-scanning direction is updated in synchronization with this. As a result, image data GI) is enlarged and interpolated in the sub-scanning direction.

In the above embodiment, the calculation of the partial interpolation coefficient 13'r+ is performed simultaneously using the current readout raster and the previous 7 lines of readout rasters, but in reality, the partial interpolation coefficient 13'r+ is calculated using the current readout raster and the preceding 7 lines of readout raster, but in reality, the partial interpolation coefficient is There are cases where it is desired to perform interpolation processing depending on image data.

In this case, for example, the adder 14-5 output and each LUTl
The phase of the current interpolation position may be delayed by inserting a four-stage shift register between 7-1 and 17-8.

The sub-scanning direction interpolation adder group 18 includes LUT17-1 to LUT
17-8 and outputs the total sum. Thus, the output of the adder 18-7 is a temporary partial interpolation value G based on the sub-scanning direction image data GD at the current interpolation position.
D.

Next, interpolation processing in the main scanning direction will be explained. The main scanning direction interpolation processing data buffer group 19 sequentially stores (latches) partial interpolation values in the sub scanning direction in synchronization with the carry signal HCAR.
do. This memory is stored from latch 19-1 to latch 19-2,
The latch 19-2 to latch 19-3 are latched in a shift register format in synchronization with the carry signal HCAH.

Therefore, if the current point is when the output of the adder 18-7 becomes valid, the latch 19-7 will hold the partial interpolation value GDY of seven columns before.
7, and latch 19-6 is 6 rows in front...
Similarly, the latch 19-1 holds the partial interpolation value GD of the previous column,
, holds.

The main scanning direction interpolation table group 20 includes eight look-up tables (LUTs) 20-1 to 20-8. Each LUT contains the corresponding partial interpolation value GDY and the real pressure 111t measured in the main scanning direction from the current sampling position.
Interpolated image data H to be generated at the interpolation position with the density of the partial interpolation value GDY using x (output of adder 15-5) as input.
The product of the contribution rate (partial interpolation coefficient Ix in the main scanning direction) indicating how much contribution should be made from the main scanning direction to the density of GD and the corresponding partial interpolation value GDY is output. For example, LUT2
0-1 is the current partial interpolation value GDyo and adder 15-
The current interpolated position Generates and outputs the product value internally. Similarly, LUT20-8 is FIFO1
Partial interpolation value GDY7 of 7 columns before 9-7 output and adder 1
The interpolation position X of 5-5 is used as an address input, so that the LUT 20-8 inputs the value of the predetermined coefficient Ix (partial interpolation coefficient) corresponding to the interpolation position x7, which is the current interpolation position X plus 7 pixel intervals. The product value of the partial interpolation value GDY is internally generated and output. In this way, the partial interpolation value GDY is sequentially updated in synchronization with the generation of the carry signal HCAR, and the product with the aforementioned partial interpolation coefficient Ixn is sequentially calculated. During this period, the pixel synchronization signal PGDS is normally generated at a rate of one or more (1, 2, 1, 3, etc.), and the interpolation position X in the main scanning direction is updated in synchronization with this. As a result, image data GD is enlarged and interpolated in the main scanning direction.

In the above embodiment, the calculation of the partial interpolation coefficient Ix is performed simultaneously by using 8 pixels of the current partial interpolation value GDY and the preceding 7 partial interpolation values GDY, but in reality, the calculation of the partial interpolation coefficient Ix is performed approximately at the current interpolation position. There are cases where it is desired to perform interpolation using partial interpolation values IY for four pixels before and after. In this case, for example, the adder 15-5 output and each LUT 20-1 to 20
The phase of the current interpolation position may be delayed by inserting a four-stage shift register between -8 and -8.

The main scanning direction interpolation adder group 21 includes LUT20-1 to LUT
20-8 and outputs the total sum. Thus, the output of adder 21-7 is the final interpolated value V at the current interpolated position. ut (interpolated image data) (GD)
It is.

FIGS. 4(A) and 4(B) are operation timing charts illustrating the manner of synchronization between the interpolation processing circuit 8 and the image printer 10. In FIG. 4(A), when the image printer 10 becomes ready, the sequence control circuit 13 outputs a read address enable signal and the image printer 10 is actually printing. Subsequently, the page synchronization signal P is sent from the image printer 10.
PDS and scanning line synchronization signal PSDS are output, and interpolation processing is performed in synchronization with these.

FIG. 4(B) is an enlarged view of one scanning line period in FIG. 4(A). In FIG. 4(B), a pixel synchronization signal PGDS is generated at predetermined intervals within one scanning line period, and a carry signal HCAR is normally generated at a constant rate smaller than this. On the other hand, the interpolated image data HGD is the pixel synchronization signal PGD.
It is output in synchronization with S.

Next, the enlarging interpolation process for the original image will be explained in detail.

FIG. 7(A) is a diagram showing the original image of a person, and FIG.
B) is a diagram showing an image obtained by enlarging and interpolating the original image of FIG. 7(A). The enlargement ratio of the image in the figure is, for example, 5.75 times in the main scanning direction and 4.25 times in the sub-scanning direction. Such image enlargement can be realized by increasing the number of pixels of the original image using the interpolation method according to the present invention.

FIG. 8(A) is a diagram showing a partial area of the original image.
Figure (B) shows the partial area of Figure 8 (A) in the main scanning direction.
．． 75 times, and 4.25 times in the sub-scanning direction. The pixel pitch in FIG. 8(A) corresponds to the data sampling pitch when the video image signal vS is sampled, and FIG. 8(B) shows a relatively enlarged portion using this sampling pitch as the minimum unit. The area is shown. Therefore, partial area a of the original image is partial area a' of the enlarged image.
The interpolation position assumed in the original image corresponds to the output pixel position b' in the enlarged partial area a'. Therefore, interpolated image data HGD is formed at the output pixel position b', but the value (density) of the interpolated image data HGD
is determined by using the values of a plurality of predetermined image data GD surrounding the output pixel position b', and by weighting the contribution rate (interpolation coefficient) determined by the distance from the pixel position b' to each image data GD. , by taking their sum.

Next, the case of determining the value of the output pixel position b' in FIG. 8(B) will be explained. This is equivalent to finding the real value corresponding to the interpolated position in FIG. 8(A). Therefore, Figure 8 (A
) is rewritten as shown in Figure 9. In FIG. 9, the position of the x mark is point b.

Here, the pixel positions of the original image are indicated by O marks, and the sampling pitch of the image data GD is equal in both the main scanning and sub-scanning directions. Also, in Figure 9, the pixel spacing of the original image is set to a unit ratio of 1! ! (For example, distance 11), and the coordinates in the main scanning direction are positive from the interpolation position of the x mark, and the right direction is positive and the left direction is negative, and the coordinates in the sub-scanning direction are also from the interpolation position of the x mark. Consider introducing another coordinate system with the origin as positive in the downward direction and negative in the upward direction. Therefore, in the new coordinate system, the position of the original image can be expressed by applying ±ΔX and ±Δy, respectively. ΔX and Δy are distances from a pixel position C of the original image having coordinate values not exceeding the interpolation position in the main scanning direction and the sub-scanning direction, respectively.

Now, let the distance in the main scanning direction between the pixel position and the interpolation position of a certain original image be X, and similarly let the distance in the sub-scanning direction be y, and then It is assumed that the contribution rate (interpolation coefficient) (y) in the scanning direction is expressed by the following equations.

πx 2 4
Here m=4. A multiple of 2 was chosen, such as n=4. Next, multiply these I (x) and I (y) by the value of the input image data GD, calculate this operation for all input image data GD within the range used for interpolation, and use the sum as the interpolated value ■. do.

That is, the value of the interpolated value I (interpolated image data HGD) is set to ■. U
7, the interpolation point is divided into surrounding 8 pixels x 8 as shown in Figure 9.
Considering that interpolation is performed using a total of 64 pixel data of the raster, VOLIT is calculated using the following formula, vouT=Σ Σ(V (i-Δ
x, j-Δy)ls-3J=-3xI(i-Δx)XI(j-Δy)). This means that after calculating all the values for each of the 64 pixels multiplied by the interpolation coefficient, the sum of the 64 values is calculated. However, attempting to realize this as is would require enormous amounts of hardware and would be impractical. Also, if it is performed using software, sufficient processing speed cannot be obtained. Therefore, the above equation for V 0LIT is modified as follows.

V2OdixI(i-Δx) Thus, first performing interpolation in the sub-scanning direction and then using the result to perform interpolation in the main scanning direction is equivalent to the above. The first embodiment described above is realized according to this method.

FIG. 10 is a diagram showing details of the interpolation principle in the main scanning direction. In the figure, the symbol (V+-ΔX) is the coordinate (i-ΔX)
This is the value of the image data GD at the point.

Based on these, the value vo of the interpolated image data HGD is obtained as follows.

Vo=Σ Vl-g ・ω5inc(i, ΔX)1
*-N/2 + 1 Also, ω5inc (i, ΔX) = ω(i, Δ)()・5inc(i, ΔX)π(i-
ΔX) = 172[1+cos(2yrc(i-ΔX))]π
(i-ΔX). The LUT instantly performs the equivalent of this calculation.

Figures 11 (A) to (C) are diagrams showing how image data increases (how the number of pixels increases) when the original image data is interpolated in the sub-scanning direction and then interpolated in the main scanning direction. . In FIG. 11(A), a scanning line interpolated only in the sub-scanning direction is created using eight consecutive scanning lines in the input image, that is, using pixels at the same position in the sub-scanning direction on each scanning line. do. In FIG. 11 CB), interpolation is performed in the main scanning direction using pixels on the scanning line interpolated only in the sub-scanning direction. This results in an output scan line with an increased number of pixels in one scan line and an increased number of scan lines in the image as shown in FIG. 11(C).

[Second Embodiment] The second embodiment relates to generalization of the selection of input pixel regions required for calculating interpolated values. That is, in the second embodiment, each correction coefficient formula I (x
), I (y) are taken as follows, and the input pixel area is expanded from the 8×8 described in the first embodiment to 2m×2n.

[Third example] The third embodiment relates to a modification of the configuration shown in FIG.

That is, the LUT groups of the interpolation table groups 17 and 20 may be configured as an LUT that outputs only a predetermined coefficient ω(x) or ω(y) and a multiplier that calculates the product of the LUT output and the pixel value. .

In this way, it is possible to significantly reduce the capacity of the LUT. Incidentally, in the third embodiment, the size can be determined by a trade-off between image quality and circuit scale, and this also shows that the present invention itself does not depend on size.

[Fourth example] The fourth embodiment relates to a modification of the configuration shown in FIG.

That is, in the first embodiment, the DDA 14.15 was configured to add and hold only the decimal part and output the carry signal CAR, but by further expanding the bit length of the ODA,
It may also be configured such that an integer part can also be held and the input pixel position can be specified using this integer part.

[Fifth example] The fifth embodiment relates to a modification of the configuration shown in FIG.

That is, in the first embodiment, the increment value and initial value of DDA14.15 are initialized by the control unit of the image printer 1o, but there are other methods as well, such as shorting out the date switch or the short circuit signal bin. You can also do this using This embodiment is shown in FIG.

[Sixth Example] The sixth embodiment relates to a modification of the configuration shown in FIG.

That is, in the first embodiment, the DDA 14.15 was configured to hold the initial value and the current value separately, and input them selectively to the adder using the selector. For example, in the case of DDA in the main scanning direction, reinitialization may be performed in synchronization with the scan line synchronization signal PSDS, and in the case of DDA in the sub-scanning direction, reinitialization may be performed in synchronization with the page synchronization signal PPDS. In this case, an initial value holding circuit and a selector are not required, but scanning such as returning an interrupt signal to the sequence control circuit by inputting a synchronizing signal or resetting the holding circuit with a synchronizing signal is required.

[Seventh Example] The seventh embodiment relates to another aspect of the interpolation coefficient formula.

That is, assuming that interpolation is preferably performed at 2m×2n points surrounding the interpolation position, the interpolation coefficient equation 1 (x), I (
y) may be as follows.

Although values can be determined more easily in the seventh embodiment than in the second embodiment, the density stripes occurring at pixel intervals are strong.

[Eighth Example] The eighth embodiment relates to another aspect of the interpolation coefficient formula.

That is, assuming that interpolation is preferably performed at 2mxZn points surrounding the interpolation position, the interpolation coefficient formulas I (x), I (y
) may be as follows.

πX
filπyn In this way, ω(x) and ω(y) can be diversified.

According to this, the occurrence of dark and light stripes is weaker than in the first embodiment.

Furthermore, if the same number of interpolation points is obtained, the sharpness of the image after processing is reduced, albeit slightly. Therefore, there is room for different uses depending on the purpose.

[Ninth Example] The ninth embodiment relates to another aspect of the interpolation coefficient formula.

That is, assuming that interpolation is preferably performed at 2m×2n points surrounding the interpolation position, the interpolation coefficient equation 1 (x), I (y
) may be as follows.

π x m
mπy
n n In this way, ω(x) and ω(y) can be diversified.

According to this, the occurrence of dark and light stripes is weaker than in the first embodiment.

In addition, the sharpness of the processed image deteriorates, albeit slightly, when performing interpolation in which the same number of interpolation points is obtained.

Therefore, there is room for different uses depending on the purpose.

[Tenth Example] The tenth example relates to another aspect of the interpolation coefficient formula. That is,
Preferably, interpolation is performed at 2mXZn points surrounding the interpolation position, and the interpolation coefficient formula I (x).

1(y) may be as follows.

π x I(, [ω
Hatake m]πy Io[ωbn1 Here, Io[] is the modified zero-order Bessel function of the first kind, and ω
1. ωb is a setting parameter, and a value in the range −〈ω,〉−mm− ωb＜−nn is selected and used. ■. [X] can be approximately determined by, for example, etc.

In this way, ω(x) and ω(y) can be diversified.

The occurrence of dark and light stripes is weaker than in the first embodiment. Furthermore, if the same number of interpolation points is obtained, the sharpness of the image after processing is reduced, albeit slightly. Therefore, there is room for different uses depending on the purpose.

[Eleventh Example] The eleventh example relates to another aspect of the interpolation coefficient formula. That is,
Assuming that interpolation is performed using 4x4 points surrounding the interpolation position, the interpolation coefficient formulas t (x) and I (y) may be as follows.

I(x)-(a+2)x'-(a+3)x"+1
(0≦IXI<1)I (x)=ax3-5ax'+8
ax-4a (1≦IXI<2)I (x)
−〇 (2≦1×1) where a is a real constant.

The same applies to I(y).

[Effects of the Invention] As described above, according to the present invention, by using an interpolation coefficient obtained by adding a second coefficient determined by the interpolation position to the 5ine function, image quality that is easy to implement and far better than that of the conventional example can be achieved. It has the effect of being able to obtain

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram of a digital image interpolation device according to a first embodiment of the present invention. FIG. 2 is a block diagram showing details of an interpolation processing circuit 8 of the first embodiment. FIG. 3 is a block diagram of a digital image interpolation device according to a first embodiment. A flowchart of the image data processing procedure by the sequence control circuit 13 of FIG.
B) is an operation timing chart explaining the synchronization mode between the interpolation processing circuit 8 and the image printer 10, FIG. 5 is a schematic diagram explaining the operation mode of the DDA of the first embodiment, and FIG. Figure 7 (A,) is a diagram showing the storage mode of image data; Figure 7 (B) is a diagram showing the original image of a person;
is a diagram showing an image obtained by enlarging and interpolating the original image in Figure 7 (A), Figure 8 (A) is a diagram showing a partial area of the original image, and Figure 8 (B
) is a diagram showing the partial area of Figure 8(A) multiplied by 5.75 times in the main scanning direction and 4.25 times in the sub-scanning direction, Figure 9 is a diagram showing details of Figure 8(A), Figure 10 is a diagram showing details of the principle of interpolation in the main scanning direction, and Figures 11 (A) to (C) are image data obtained when interpolation is performed in the main scanning direction after interpolation in the sub-scanning direction using original image data. 12 is a block diagram showing a part of the configuration of the DDA of the fifth embodiment. FIG. 13 is a diagram showing tables 17 and 20 of the embodiment. be. In the figure, 1... Video signal source, 2... Luminance signal extraction circuit, 3... Synchronization signal extraction circuit, 4... Sampling cycle generation circuit, 5... A/D conversion circuit, 6... ... γ conversion circuit, 7... Frame memory 7.8... Interpolation processing circuit, 9... Printer I/F circuit, 10... Image printer, 11... Write address control circuit, 12 ...
Read address control circuit, 13... Sequence control circuit. Patent ratio - applicant Canon Co., Ltd.

Claims (5)

[Claims] (1) A digital image interpolation method for providing interpolated image data to an image output device, comprising: storing a plurality of image data in integer coordinates of orthogonal coordinates X and Y; and real number coordinates of the orthogonal coordinates X and Y. a step of setting an arbitrary interpolation position, a step of selecting 2m x 2n pieces of image data so as to surround the interpolation position with the set interpolation position as the approximate center; distance x
, y, and calculating predetermined coefficients ω(x), ω(y
), and calculate the interpolation coefficient I from the determined distances x, y and the coefficients ω(x), ω(y) using the following formula: I=ω(x)・(sinπx/πx)・ω (y)・(s
inπy/πx), and calculating each product of the interpolation coefficient I obtained above and the image data corresponding to the interpolation coefficient, and obtaining interpolated image data by the sum of the obtained 2m×2n products. 1. A method for interpolating a digital image, comprising the step of forming a . (2) The values of coefficients ω(x) and ω(y) are expressed as follows, ω(x)=
1/2(1+cos[πx/m])ω(y)=1/2(
1+cos [πy/n''). (3) The digital image interpolation method according to claim 2, characterized in that m=n=8. (4) Interpolated image data is formed by inputting temporary interpolated image data obtained only for distance x or y, and then
2. A digital image interpolation method according to claim 1, characterized in that a calculation is performed for a digital image. (5) A digital image interpolation device for providing interpolated image data to an image output device, comprising: an image memory that stores a plurality of image data at integer addresses of orthogonal addresses X and Y; and real numbers of the orthogonal addresses X and Y. interpolation position setting means for setting an arbitrary interpolation position in the address; image data selection means for selecting 2m×2n pieces of image data so as to surround the interpolation position with the set interpolation position as the approximate center; distance calculating means for calculating distances x, y from the interpolation position to the selected image data; and predetermined coefficients ω(x), ω(y) from the calculated distances x, y.
), and a coefficient calculation means for calculating the value of the distance x, y and the coefficients ω(x), ω(y) as follows: I=ω(x)・(sinπx/πx)・ω(y)・(s
(inπy/πy); and the product of the interpolation coefficient I obtained above and the image data corresponding to the interpolation coefficient, and the sum of the obtained 2m×2n products. A digital image interpolation device comprising an interpolation image data forming means for forming interpolation image data.