JPH10198338A - Image processing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing device for converting image data with a less memory capacity. SOLUTION: In such a case that image data having 640×480 dots is converted into image data having 1280×960 dots, a memory 10 for processing 640×16 lines is prepared. Initially, actual data for eight lines are developed in the upper half parts of the memory 10 and are processed, and then the data are developed in the lower half parts of the memory and rare processed. With the repetitions of these steps, four bits from the most significant bits of the coordinate values of four actual data calculated from desired dots on virtual coordinates are used as line numbers ('0' to '15') of the actual data developed in the memory 10.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an image processing apparatus for converting the size of image data using a small memory area.

[0002]

2. Description of the Related Art In recent years, digital still cameras have become widespread. In this digital still camera, photographed image data is stored as digital data in a storage device. The image data is displayed on an external or built-in display device, or printed by a printing device connected via an interface, and can be viewed at any time. The demand for a printer as the print output device is increasing.

The printer has a control device having a data memory, a work memory, a printing memory, etc., and controlling the entire printer such as communication control with a camera, image processing, and printing control. The printer uses a serial I / O
It is connected to a digital still camera (hereinafter simply referred to as a camera) via an F (interface). Generally, image data transmitted from a camera is compressed by JPEG or the like, and the compressed image data is transmitted to the printer in the same format. The printer temporarily stores the compressed image data in the data memory and performs image processing to convert the compressed image data into print data. One of such image processing techniques is image size conversion. The image size conversion means, for example, converting image data of 640 × 480 dots into 1280 × 9
This is to convert the image data into 60 dot image data. The conventional image size conversion is based on the premise that image data for one frame is expanded in a memory, reads image data from the memory, performs an image size conversion operation,
Writing to another memory area (for a printer, for example, a printing memory) is performed.

FIG. 15 is a diagram schematically showing, for example, an area of a frame memory of 640 × 480 dots in which image data is expanded when such image processing is performed. The conventional image processing will be further described with reference to FIG. First, in order to represent an arbitrary dot existing in the frame memory, the X axis is set in the 640 dot direction of the frame memory, and the Y axis is set in the 480 dot direction. Hereinafter, a system based on these coordinate axes will be referred to as real coordinates. As a result, an arbitrary dot (pixel data) existing on the frame memory can be represented as one point on the real coordinate system. Hereinafter, the pixel data existing on the frame memory will be referred to as actual data.

When the above-mentioned 640 × 480 dot image data is converted into an enlarged image size like 1280 × 960 dot image data, the value between the actual data of the original image data and the actual data is enlarged. It must be obtained by interpolation as new data. In other words, it is necessary to perform a process of obtaining the value of a dot at an arbitrary position by interpolation. This interpolation operation is performed by the X mark K in FIG.
Taking the case of finding the value of 1 as an example, generally, four points of actual data surrounding the X mark K1, that is, the upper left dot J1, the dot J2 on the right, and the dot J one line below the dot J1
3. Calculate based on the value of each dot J4 on the right. This calculation method is well known. Then, for the interpolation operation, first, the position of the mark K1 in the figure must be indicated by coordinates.

FIG. 16 is a diagram showing a coordinate system newly introduced to represent the position of new data (hereinafter referred to as a desired dot) to be obtained by interpolation as one point on the coordinate system. The figure shows that the new coordinate system is divided into r equal parts (r = 4) between the respective real data of the real coordinate system of FIG. 15 and the X-axis direction (640 dots) in the case of the real coordinates is from 2559 to 0 to 2588. In the same manner, dots are represented in the Y-axis direction (480 dots) from 0 to 191.
8 is represented by 1919 dots.
In this case, the desired dot can be handled as one point on a new coordinate system. In the new coordinate system, the actual data takes a value obtained by multiplying the coordinate value in the real coordinate system by r. Hereinafter, the new coordinate system will be referred to as virtual coordinates. In the following description, the virtual coordinate system is defined as <X, Y
> Such as <> or uppercase, and the real coordinate system is (x,
() or lower case like y).

In the above new coordinate system, r = 4. However, if the desired coordinates cannot be represented by dividing the actual coordinates into four, r is set to a larger value. It is sufficient to divide the real dots into finer parts than the above and set a new coordinate system that can correctly represent the desired dots.

After setting the virtual coordinate system in this way, interpolation calculation is performed using actual data of four points around the desired dot to calculate the value of the desired dot. Virtual coordinate value of desired dot <X,
In order to obtain the actual data of the four surrounding points from Y>, coordinate values are calculated by quotients obtained by dividing the values of the virtual coordinate values <X, Y> by r. The coordinate values obtained by this calculation are the actual coordinate values of the actual data at the upper left of the desired dot. Assuming that these coordinate values are (x ₀ , y ₀ ), the actual coordinate values of the remaining three points around the desired dot are (x ₀ +1, y ₀ ), (x
_0, y ₀ +1), is obtained as _{(x 0 + 1, y 0} +1). Incidentally, in the example of FIG. 16, the virtual coordinate value of the x mark position is <
3, 2>, and the value obtained by equally dividing the actual coordinates is r = 4, so the above calculation result is (x ₀ , y ₀ ) = (0, 0).

Next, the remainder of the above division is obtained. Assuming that the remainder obtained by dividing the coordinate value x by r is mod_x, and the remainder obtained by dividing the coordinate value y by r is mod_y, since r = 4, mod_x = 3 and mod_y = 2. This value is X, between the desired dot and the upper left dot of the desired dot.
The distance in the Y-axis direction is shown. The scale of this distance is a value on virtual coordinates. By using these, the value of the desired dot can be calculated.

That is, now (x ₀ , y ₀ ) = D ₀ ,
(X ₀ +1, y ₀ ) = D ₁ , (x ₀ , y ₀ +1) =
D ₂ , (x ₀ +1, y ₀ +1) = D ₃ (D ₀ , D ₁ , D
When the positional relationship of _2, D ₃ is reference to FIG. 16) that, _{first, D 01 = (D 1 ×} mod_x + D 0 × (r-mod_x)) / r ... (1) a to total first. _{Then, D 23 = (D 3 ×} mod_x + D 2 × (r-mod_x)) / r ... compute the (2). Then, the desired dot D _{0123 is obtained} from D ₀₁ and D _23.
Can be obtained by calculating D ₀₁₂₃ = (D ₂₃ × mod_y + D ₀₁ × (r-mod_y) / r (3)).

By performing this for all necessary dots, the size conversion operation is completed. When actually performing the size conversion operation, the calculation is started from the upper left dot, the value of the virtual coordinate Y is first fixed, and the calculation is performed while the virtual coordinate value is advanced in a certain step in the X-axis direction. When the axial direction ends, the virtual coordinate value is advanced by one step in the Y-axis direction, and the calculation in the X-axis direction is repeated again.

For example, from the virtual coordinates in FIG.
When obtaining an image of 960 dots, a step in the X direction (hereinafter, referred to as an interpolation step) and a step in the Y direction (hereinafter, referred to as a Y interpolation step) are each set to 2 and the virtual coordinates <0, 0> are shifted in the X-axis direction. <2,0>, <4,0
, ..., <2558, 0>, and the next line is <0, 2>, <2, 2>, <4, 2>, ..., <25
58, 2>, and sequentially calculate this until the last line <
0,1918>, <2,1918>, <4,1918
, ..., <2558, 1918> in the order of 9
60 × 1280 image data is obtained.

Note that the rightmost <2558,0>, <255
, <2558, 1918>, (x ₀ +1, y ₀ ), (x ₀ , y ₀ +1), (x ₀
+1, y ₀ +1) are (640, ₀ ), (640, 4
80), and no actual data exists. In that case, the calculation is performed by using a predetermined value (for example, 0) instead of the actual data.

FIG. 17 is a flowchart showing the operation of image processing in a conventional printer. The image processing operation will be briefly described with reference to this flowchart. First, when a print instruction is issued by a user's operation of a print start key or the like (step S1), the image processing apparatus of the printer communicates with a connected camera (step S1).
2), receive JPEG compressed data of an image to be printed (here, an image of 640 × 480 dots) and store the received data in a data memory (step S)
3).

Then, JPEG expansion for 16 lines is performed (step S4). This processing is generally performed by JPE
The expansion process of the G compressed data is performed in units of 16 lines in accordance with the generation of the expanded data in units of 16 × 16 dots. The JPEG decompression process is performed using the JPEG decompressed data area of the work memory, and Y (luminance), Cb, and Cr data for 16 lines are generated.

Next, for the 16 lines, YCb
The conversion from the Cr data to the RGB data is performed (step S5), and the color to be printed is yellow (Yellow).
w), magenta or cyan (Cya)
It is determined which of n) the colors are (steps S6 and S7). If the color to be printed is yellow, color conversion for yellow of the 16 lines of RGB data is performed (step S8). If the color to be printed is magenta, color conversion for magenta of the 16 lines of RGB data is performed. If the color to be printed is neither yellow nor magenta (step S9), color conversion for 16 lines of RGB data for cyan is performed (step S10).

The color-converted data is stored in the YMC data area of the work memory (step S1).
1). Thereafter, it is determined whether or not data expansion of all lines (480 lines in this example) has been completed (step S).
12) If not completed, the above steps S4 to S11
repeat. When the data expansion of all the lines, that is, 480 lines is completed (Y in S12), all the data for one screen is sequentially read from the YMC data area of the work memory, and the size is converted according to the resolution of the printer. The print data after the conversion processing is developed in the print memory (step S13). When the development is completed, printing of the developed color is performed (step S1).
4). This printing process is performed by reading from a printing memory by a print control circuit (not shown). When printing is completed, the same processing is repeated, and printing of all colors is completed.

[0018]

By the way, in the above-mentioned image processing method, a capacity of 64 bytes × 480 = 307200 bytes (one dot is converted to one byte) is used as a work memory area for size conversion. ing. The capacity of the memory temporarily used for such a work is originally small. Otherwise,
This hinders the realization of cost reduction and the miniaturization of the apparatus.

An object of the present invention is to provide an image processing apparatus for performing image data conversion with a small memory capacity.

[0020]

According to the first aspect of the present invention, there is provided an image processing apparatus, wherein the number of lines of a memory used for image processing is two.
ⁿ (n = 1, 2, 3,...). The memory is configured such that the address value is represented by a continuous numerical value, for example.

According to the image processing apparatus of the present invention, the coordinates (x, y) of each pixel on the memory used for processing the image by treating the original image data as a two-dimensional plane (m , N), and when the number of lines in the memory is set to 2 ⁿ (n = 1, 2, 3,...), When calculating the address on the memory, the lower n of the coordinates
It is configured to perform processing using bits.

The image processing is, for example, an interpolation calculation for converting the size of the image data. In addition, the interpolation calculation is configured to be performed while determining whether or not the interpolation calculation is possible by a predetermined conditional expression, for example, as described in claim 5. For example, the conditional expression is defined as follows: when the virtual coordinates have a resolution r times the real coordinates and the number of times of real data development for each half of the number of memory lines is i, The virtual coordinate value Y of the target dot for the interpolation calculation is “Y <(2 ⁿ⁻¹ × (i + 1)”
−1) × r ”or the real coordinate value y of the dot of the real data at the upper left of the target dot is“ y <(i × 2 ⁿ⁻¹ +2)
^{When n-1} -1) is satisfied, it is determined that interpolation calculation is possible. Then, for example, as set forth in claim 7, the determination based on the conditional expression is not performed at the time of the final processing.

According to an eighth aspect of the present invention, in the interpolation processing for obtaining an arbitrary point in the space from known points around the space, it is determined whether or not the calculation performed in the processing can be omitted. , When it can be omitted, the calculation is controlled not to be performed.

[0024]

Embodiments of the present invention will be described below with reference to the drawings. First, in the first embodiment, when decompressing input image data (hereinafter, also simply referred to as an input image) input by the camera 3 and compressed by JPEG or the like, size conversion is performed in parallel with decompression calculation. To perform an interpolation operation. The configuration of the work memory used for the interpolation operation is determined by the data configuration unit (block unit of 16 × 16 dots) generated in the data expansion process of JPEG.
, The configuration is for 16 lines of image data. Then, 16 lines of image data (the source image is 6
When 40 × 480 dots are generated, every time 640 × 16 dots of data are generated, an interpolation operation for size conversion is performed, thereby reducing the memory capacity required for the size conversion operation.

FIG. 1 is a block diagram showing the configuration of the image processing apparatus according to the first embodiment. The image processing apparatus 1 shown in FIG. 1 is, for example, an interface controller of a printer, and the print control circuit 2 is, for example, an engine controller of the printer. The above image processing apparatus 1
, A digital camera (hereinafter simply referred to as a camera) 3 as an external host device is connected via a serial interface (I / F).

The image processing apparatus 1 has a CPU 4 and a CP
In U4, a print start key 5 for the user to instruct the start of printing, a data memory 6 for storing image data input from the camera 3, a JPEG decompression process of the image data, and a size conversion in parallel with the decompression process A work memory 7 for temporarily using the interpolation calculation for the operation, a printing memory 8 for storing one screen of print image data after size conversion, a ROM 9 for storing processing programs, coefficients, and the like are connected. ing. The print control circuit 2 includes a print memory 8
The printer engine is controlled based on the print image data read from the printer, and printing is executed.

FIG. 2 shows an address map of a work memory area (hereinafter, referred to as a YMC work area) for yellow (magenta, cyan) used for data expansion (JPEG expansion) and size conversion used in the present embodiment. I have. The YMC work area 10 shown in FIG.
Image data of 40 × 16 dots are line-sequentially arranged from address 0 to address 10239. The YMC work area 10 includes a YMC work area A11 (hereinafter, simply referred to as area A11) and a YMC work area B12.
(Hereinafter, simply referred to as area B12). Each of these two areas has a capacity of 640 dots × 8 lines, and the addresses of image data in both areas are continuous. That is, the area A11
640 × 8 dots are addresses “0” to “51”.
19 "and the next address" 5120 "
From area B12 to address "10239"
× 8 dots are line-sequentially continuous. As described later in detail, the decompressed data is stored in two areas A11 every eight lines.
And B12 alternately.

FIG. 3 is a diagram showing the relationship between the address of real data and the actual coordinates to be developed in the YMC work area 10 and the state of development of the real data. The white circles developed in the lower half of the figure are indefinite data in the initial state, and the black circles developed in the upper half of the figure are actual data just developed.

As shown in FIG. 3, each dot of the YMC work area 10 has addresses "0" to "10" on the memory.
239 "(in the figure, each data is indicated by []), and on the secondary plane are represented by real coordinates (0, 0) to (2556, 60) as in the case of FIG. In the figure, each data is indicated by ().)

FIGS. 4, 5 and 6 show actual data and virtual coordinates (<> in each figure, respectively) that are successively developed every eight lines in the YMC work area 10 following FIG.
(Indicated by ()) and a development state of actual data. The black circles shown in FIGS. 4 and 5 are the actual data just developed at the processing timings shown in the figures, and the mesh circles are the actual data developed at the previous processing timings.

In the present embodiment, the position of the real data is represented using a virtual coordinate system in which the real data is divided into four equal parts as in the case of FIG. FIG. 4, as described above,
3 shows a state immediately after the next eight lines of data have been developed in the area B12 after the data development shown in FIG. 3, and the virtual coordinate values shown below the real data just developed are the same as those in the previous processing timing. It is continuous with the virtual coordinate values of the actual data expanded above.

Then, after the data expansion shown in FIG.
In FIG. 5 showing a state in which the data of the next eight lines has been developed in the area A11, the virtual coordinate values shown below the real data just developed are the actual data developed below in the previous processing timing of FIG. Are continuous with the virtual coordinate values of. This is the coordinate value of the virtual coordinate system when it is assumed that the data has been expanded in the memory of one frame as in the case of FIG. Similarly, in FIG. 6, the virtual coordinate values shown below the real data just developed in the lower area are continuous with the virtual coordinate values of the real data developed above at the previous processing timing in FIG. 5.

FIG. 7 is a diagram obtained by redrawing the actual data development diagram of FIG. 3 with virtual coordinates as shown in FIGS. 4 to 6 for use in the following description. As described above, FIGS.
As shown in FIG. 6, in the structure of the image data which is sequentially and alternately developed in the area A11 and the area B12 of the YMC work area 10 for every eight lines, the processing of performing the expansion processing and the size conversion in parallel will be described below. I do.

FIG. 8 is a flowchart showing the operation of the processing in the first embodiment. In the flowchart of FIG. 17, the processing of steps S21 to S24 is the same as the processing from the start of the processing of steps S1 to S4 to the JPEG decompression for 16 lines shown in FIG.

In the present embodiment, following the above,
First, YCbCr for upper 8 lines of 16 lines
Data is converted to RGB data (step S2).
5) Next, the color to be printed is yellow (Yellow).
w), magenta or cyan (Cya)
It is determined which of n) the colors are (steps S26 and S27). If the color to be printed is yellow (Y in S26), the RGB data for the eight lines is converted into yellow (step S28). still,
If the color to be printed is magenta (S26 is N, S27
Y), the color conversion of the RGB data for the eight lines to magenta is performed (step S29), and the color to be printed is neither yellow nor magenta (S26 is N, S27
N), that is, if cyan, eight lines of RGB
The color conversion of the data to cyan is performed (step S30).

Then, the color-converted data (actual data) is stored in an area A11 of the YMC work area 10 (step S31). A development process to the memory 8 is performed (step S32), followed by step S3.
The processing of 3 to S40 is processing for the data of the lower 8 lines, and only the memory area to be expanded is changed to the area B12, and the upper 8 of the above-mentioned steps S25 to S33.
The processing is the same as the processing for the line data.

Subsequent to the above processing, it is determined whether or not data expansion of all lines (480 lines in this example) has been completed (step S41).
N), and repeat the above steps S24 to S41. When it is confirmed that the data expansion of all lines, that is, 480 lines is completed (Y in S41), all the data for one screen is sequentially read from the printing memory 8 and one color printing is performed (step S42). The printing process is performed by the print control circuit 2 reading out the print data from the print memory 8. When printing is completed, it is determined whether or not printing of all colors is completed (step S43). If not completed (N in S43), steps S24 to S42 are repeated again to complete printing of all colors. (S43 is Y).

FIGS. 9 and 10 show the above-mentioned step S32.
It is a flowchart explaining the process in (or S40) in more detail. Here, the capacity of the YMC work area 10 is (2 ⁿ lines) × (h dots), the virtual coordinates are r = 2 ^d (d is a positive integer of 1 or more), and the resolution is r times the real coordinates. Set to have. The settings of the YMC work area 10 shown in FIGS. 3 and 7 and FIGS. 4 to 6 are obtained by setting the above coefficients to n = 4, h = 640, d =
2 is shown. Hereinafter, FIGS. 9 and 1
3 and FIGS. 7 and 4 to FIG.
Referring to FIG. 6, the size of the image data is 640 × 4
The size conversion processing operation will be further described, taking as an example the case of converting from 80 dots to 1280 × 960 dots.

In the following description, the variable KASO_
X is the virtual coordinate value (KASO_X, KAS) of the desired dot.
O_Y) and the variable KASO_Y
Represents the Y coordinate value in the same manner. Also, the variable S
TEP_X represents an X interpolation step, and a variable STEP_Y represents a Y interpolation step.
KASO_X >> d indicates that KASO_X is right-shifted by d bits, and KASO_Y >> d
Represents that KASO_Y is shifted right by d bits.

The right shift in the data (bit string)
Is to perform the division, for example, when 1-bit right shift data a (bit string a) obtained quotient of the division "a / 2", when 2-bit right shift "a / 2 ^2" quotient of, 3
Shifting right by bits gives the quotient of "a / 2 ³ ",... Shifting right by d bits gives the quotient of "a / 2 ^d ".

In the flowchart shown in FIG. 9, first, a variable KASO_Y for virtual coordinates and a counter i are initialized (step S50). In this process, KASO
The values set in _X and KASO_Y are the virtual coordinate values of the dot on which the interpolation calculation is performed first. For example, FIGS.
6, virtual coordinate values (0, 0) are set.
Therefore, the variable KASO_Y is set to “0”. “0” is set in the counter i.

Next, JPEG expansion for 2 n lines is performed to store YC in the JPEG work area of the work memory 7.
The bCr data is generated (step S51). The 2 n -th power line in this processing is, for example, as shown in FIGS.
In the example, n = 4, so in this case 2 ⁿ
= 2 ⁴ , that is, 16 lines.

Subsequently, the 2 "n-1" -th power lines are color-converted into yellow (or magenta or cyan), and the color-converted data is stored in an area A1 of the YMC work area 10.
1 (step S52). Thereby, the actual data is developed in the area A11 as shown in FIG. 7, for example.

Subsequently, it is determined whether or not the interpolation calculation of the KASO_Y line is possible (step S53). In this embodiment, since the data area of the YMC work area 10 is not a data area of one frame as in the conventional example but a data area of 16 lines, the virtual coordinates are introduced. , In the X-axis direction, but there is a problem in the Y-axis direction. This is a determination process for avoiding this.

In this process, a certain condition is set,
The interpolation calculation is performed up to the range satisfying the condition. That is, up to the 0th to 7th lines in FIG.
When a size conversion operation (interpolation calculation) is performed using actual data, the dot position at which the calculation is possible, that is, the condition of being located above the last line of the expanded real dots, is that the virtual coordinate value Y is (8 -1) × r = 7 × 4 = less than 28. In the case of 28 or more, the data of the 8th line and thereafter is used for the interpolation calculation, and the calculation cannot be performed as it is.

Therefore, when calculated according to this condition, X
The interpolation step and the Y interpolation step are each set to 2, and Y
= 28, that is, the virtual coordinates <0, 0>, <
2,0>, <4,0>, ..., <2558,0>, <
0, 2>, <2, 2>,..., 2558, 26>.

A general formula showing the above-mentioned dot conditions that can be calculated is shown below. In other words, when the YMC work area 10 has 2 ⁿ lines and the number of developments for 2 ^n-1 lines is i-th counting from 0, the condition that can be calculated in this processing is the virtual coordinate value of the desired dot. Satisfies the virtual coordinate value Y <(2 ⁿ⁻¹ × (i + 1) −1) × r (4), or the real coordinate y of the upper left real data dot of the desired dot is the real coordinate y <(i × 2 ⁿ⁻¹ +2 ⁿ⁻¹ −1) (5) The calculation formula for the discrimination in the above step S53 is based on the above formula (4). When the above conditional expression is satisfied (step S53)
Y), a variable KASO_X for virtual coordinate conversion is initialized (step S54). In this process, as described above, “0” is set to the variable KASO_X.

Subsequently, a dot whose address on the YMC work area 10 is “(lower n bits of a quotient obtained by dividing KASO_Y by r) × h + (a quotient obtained by dividing KASO_X by r) (real data at the upper left of the desired dot) , Dot J1 in FIG.
15), a dot on the right of the dot (see dot J2 in FIG. 15, address is the above address + 1), a lower dot (see dot J3 in FIG. 15, address is the above address + h), and a diagonally lower right (Refer to dot J4 in FIG. 15, the address is the above address + h + 1) for a total of four dots (step S55). The calculation of the position of the upper left real data in this process will be described later. This interpolation calculation is performed by a known method.

Thereafter, it is determined whether or not the interpolation data for one line has been created (step S56). If the creation has not been completed (N in S56), KASO_X = KAS
The virtual coordinate is advanced by one step in the X-axis direction by performing the calculation of O_X + STEP_X (step S57), and the above steps S55 and S56 are repeated. Thus, the interpolation calculation proceeds in the X-axis direction of the virtual coordinates.

When the creation of the interpolation data for one line is completed (Y in S56), it is determined whether the creation of the interpolation data for all the lines is completed (step S5).
8) If not completed (N in S58), KASO_
By performing the calculation of Y = KASO_Y + STEP_Y, the virtual coordinates are advanced by one step in the Y-axis direction (step S59),
Steps S53 and S58 are repeated. This allows
In the case of the example of FIG. 7, the processing of the interpolation calculation for the 0 to 6 lines proceeds.

Then, in step S53, the next desired dot is equal to or lower than the virtual coordinates of the real data of the seventh line (the last line of the expanded real data). If it is determined that the interpolation calculation is not possible (N in S53), the process moves to the flowchart shown in FIG.

In the process of FIG. 10, first, the counter i is incremented by "1" (step S50). Thus, the data conversion for each half area of the YMC work area 10 (in this example, for every eight lines) is sequentially counted.

Next, the color of the 2 <n-1> power lines is converted to yellow (or magenta or cyan), and the color-converted data is stored in the area B12 of the YMC work area 10.
(Step S61). This process is a process of color-converting the remaining half of the 16 lines of actual data JPEG-decompressed in step S51 and developing the data under the YMC work area 10. Thereby, as shown in FIG. 4, for example, the second line (6
Following the actual data that has been developed upward before the processing is completed up to line (line), new actual data is developed from the eighth line to the fifteenth line.

The subsequent steps S62 to S68 are the same as steps S53 to S59 in FIG. 9 except that the dots to be processed are based on actual data developed in the area B12. More specifically, the conditions that can be calculated in step S62 are as follows: From the above-described formula (4), the virtual coordinate value Y = 8 × r + (8−1) × r
= 60 (less than). Thereafter, the above steps S62 to S68 are repeated until the above condition is satisfied, that is, <0, 28>,.
.., interpolation calculation of desired dots up to <2558, 58> is performed (see virtual coordinate values of real dots on the 7th to 15th lines in FIG. 4).

Here, as shown in FIGS. 7 and 4, the actual data which is first developed twice every eight lines is YM
The line number in the C work area 10 matches the line number for one frame. That is, the actual data position on the virtual coordinates on the YMC work area 10 matches the actual data position on the virtual coordinates for one frame. Therefore, there is no problem in the operation of calculating the actual data position based on the position of the desired data on the virtual coordinates in step S55 or step S64.

When the interpolation calculation of the KASO_Y line becomes impossible (the last line and the fifteenth line in FIG. 4) in step S62 in FIG.
Returning to step S51, a new 16-line JPEG
After the decompression, in step S52, as shown in FIG. 5, the actual data developed in the area A11 includes the coordinates <0, 60>,.
2558, 90>, the virtual coordinates <0, 64>,
, <2256, 92>, the virtual coordinates of this actual data on the YMC work area 10 are the same as those in FIG. That is, the coordinate value KAS of the actual data
O_X always matches the virtual coordinate value of the YMC work area 10, but the coordinate value KASO_Y does not match the virtual coordinate value of the YMC work area 10 after exceeding 15 lines. Thus, the position on the virtual coordinates of the YMC work area 10 does not match the position on the virtual coordinates for one frame. In this state, after that, as shown in FIG. 6, the actual data of the next eight lines is developed in the area B12, and <0,92
,..., <2558, 122>. This also applies to actual data that is repeatedly developed in the area A11 and the area B12 sequentially for every eight lines thereafter.

However, the present invention does not cause any trouble even in such a case. Rather, the four YMC work areas 1 surrounding the virtual coordinates of the desired dot (position data on the virtual coordinates for one frame) can be obtained very easily.
The actual data position on 0 can be calculated. Hereinafter, this will be described with reference to FIG.

Now, it is assumed that the desired dot is an X mark "virtual coordinates <2, 106>" in FIG. First, <X, Y> is r
The quotient when divided by (in this case, r = 4) is obtained. Dividing by 4 is obtained by shifting right by 2 bits as described above. The virtual coordinates of the desired dot described above <
By shifting the coordinate values “2” and “106” of “2, 106>” to the right by 2 bits, (0, 26) is obtained. Therefore, the interpolation calculation includes the actual coordinates (0, 26) of the upper left dot, the actual coordinates (1, 26) of the upper right dot, and the actual coordinates (0, 2) of the lower left dot.
7), the operation is performed using the actual coordinates (1, 27) of the lower right dot.

Incidentally, an address on the YMC work area 10 must be calculated from the actual coordinate values. As described above, the real coordinate value x has a one-to-one correspondence with the data on the YMC work area 10, but the real coordinate value y does not correspond (that is, the virtual coordinates do not correspond). still,
7 and 4, the line number on the YMC work area 10 coincides with the value of the Y axis of the actual coordinates.

Therefore, focusing on the lower 4 bits of the real coordinate y, the lower 4 bits of the real coordinate y = 26
0H, that is, 10 in decimal. Similarly, the lower 4 bits of the real coordinate value y = 27 are 11 in decimal. By the way, these values are the line number on the YMC work area 10 where the data 15 and 16 of the real coordinate value y = 26 are developed and the YMC work where the data 17 and 18 of the real coordinate value y = 27 are developed. Each matches the line number on the area. That is, as can be seen from FIG. 6, the actual coordinate value y =
26 are present on the tenth line when the number of lines in the YMC work area 10 is counted from 0, and the real coordinate data 17, 18 and the like having the actual coordinate value y = 27 are the YMC work area. It is present on the eleventh line of area 10.

Thus, by determining the value of the lower 4 bits of the actual coordinate value y, the position on the YMC work area 10 where the actual data of the actual coordinate value y exists can be easily determined. In other words, the actual coordinates (0, 26), (1, 26), (0, 27), (1, 2)
7) Actual data on the YMC work area 10 can be obtained, and thereby interpolation calculation can be performed.

By the way, the lower 4 bits of the real coordinate value y and Y
The number of lines on the MC work area is the same
This is because the number of lines in the work area 10 is a power of two. Assuming that the number of lines in the YHC work area 10 is 2 ⁿ (n = 1, 2, 3,...), The lower n bits of the real coordinate y and the number of lines in the YMC work area 10 can be matched.

Thus, the address calculation for reading the actual data from the YMC work area 10 is simplified.
In other words, the calculation is simplified by setting r to a value of a power of two. However, this is 2 in the YMC work area 10.
^After developing ^n-1 lines (8 lines in this example), the condition is that the calculation should be performed to the maximum possible range at that time.

It should be noted that only interpolation processing for the lower 2 ^{n -1} lines based on the final decompressed data (actual data) in a certain image does not make a judgment based on this condition, but performs all interpolation calculations. For example, in the size conversion processing when the YMC work area 10 has 16 lines and the source image is 640 × 480, i at the time of development of the last 8 lines is i = 5
It becomes 9. When this is put into (i × 2 ^n-1 +2 ^n-1 -1) and calculated, it is 479. In other words, if the above determination is made during the processing for the last eight lines, the actual coordinate y
= 479 can only be calculated, causing inconvenience. To avoid this, no decision is made.

At the time of initialization, the variable KASO_X
And the values set in KASO_Y are calculated by inputting the virtual coordinate values of the dots for which the interpolation calculation is performed first. Therefore, by appropriately setting the initial value, the interpolation step, and the image size after the size conversion, the value of the image is calculated. The part can be easily enlarged. For example, in the case of the virtual coordinates in FIGS. 4 to 7, the initial values are X = 1280, Y = 960, the X interpolation step is 2, the Y interpolation step is 2, and the image size after size conversion is 640 × 480 dots. Then, after the interpolation calculation, it is possible to obtain an image of the same size as the source image in which the range of the lower right quarter of the source image is enlarged (see FIG. 11).

Further, if the X interpolation step and the Y interpolation step are set and calculated at a ratio of, for example, 1: 2, a horizontally elongated image is obtained. As described above, various image processing can be performed by combining the initial value, the interpolation step, and the image size after the size conversion.

In the first embodiment, first, the result of JPEG expansion for 16 lines is expanded on the JPEG work area, and then the converted YMC data is expanded on the YMC work area. In contrast,
The YMC work area A may be shared with the upper eight lines of the Y (luminance) signal area in the JPEG work area. That is, JPEG
At the time of decompression, Y (luminance) data for the upper 8 lines is YMC
The data is expanded to the work area A, and the lower eight lines are expanded to the JPEG work area. And YCbCr → RGB →
As for the YMC conversion, the upper eight lines are converted so as to be overwritten on the YMC work area A, and the lower eight lines are the same as in the above embodiment. In this case, the JPEG work area has only eight lines for the Y (luminance) data, and the required work memory can be further reduced.

Incidentally, the calculation of the image size conversion is very large since the interpolation calculation is performed for all the dots. If this is to be processed by the CPU, a great deal of time is required for the calculation. By the way, as described in the first embodiment as an example, 640
In the case of converting × 480 dots into 1920 × 960 dots, dots in the source image and dots in the converted image may overlap (become the same value).
In the case of the above example, (0,0) and <0,0>, (1,
0) and <4,0>,..., (X, y) and <rxx, r
× y>,..., (639,479) and <2556,1
916> corresponds to this. These dots are the remainder mod_x, mo obtained by dividing the virtual coordinate value by r during the interpolation calculation.
Both the values of d_y become 0. Similarly, equations (1) to (3)
Is calculated, mod_x or mod_y is 0.
It may be. In such a case, equations (1) to (3)
Calculation becomes unnecessary. However, in general, interpolation calculation is also performed on such data through one processing flow.

According to the present invention, in such a case, the interpolation calculation is controlled so as not to be performed, and the interpolation calculation is speeded up. Hereinafter, this will be described as a second embodiment. FIG. 12 is a flowchart of the interpolation calculation process according to the second embodiment. Further, among the variables used in the following drawings that are the same as those in the first embodiment, the meanings of the variables and the like are the same as those in the first embodiment. still,
FIG. 12 assumes use as a subroutine.

FIG. 13 is a diagram showing the input / output relationship of the subroutine. The input is the actual coordinate value (KASO_X >> d, KASO_Y >>) of the upper left actual data of the desired dot.
d), and mod_x and mod_y that are information on the distance between the desired dot and the actual coordinate value. The output is the value of the desired dot.

FIG. 14 is a positional relationship diagram of dots (actual data) D ₀ , D ₁ , D ₂ and D ₃ for explaining the processing in the present embodiment. The processing of the subroutine will be described below with reference to FIGS.

When the subroutine is called, first, the upper left dot of the desired dot is read (step S70).
Thus, data D ₀ in FIG. 14 is read. Next, it is determined whether mod_x is 0 (step S7).
1). In this determination, when mod_x is 0 (S71 is Y), a case where the desired dots are on the vertical line of data D ₀ and the data D _2, therefore, interpolation data D ₀ and the data D ₁ is not required So, move on to the next flow,
It is determined whether or not mod_y is 0 (step S72).

In this determination, if mod_y is also 0 (S7)
2 is Y), and the desired dot is on the horizontal line of data D ₀ and data D ₁ , that is, in this case, data D ₀ itself is the desired dot, and in this case, no interpolation calculation is required. the D ₀ is returned as the data of the desired dot.

[0074] In step S72, mod_y is not the 0 (S72 is N), reads out the data D ₂ (step S74), the data D ₀ and the data D _2, and mod
The calculation of the value of the virtual data D ₀₂ by using the _y (step S75). Then, the virtual data _D02 is returned as desired dot data.

Further, in the determination in step S71, mod_
If x is not 0 (the S71 N), reads out the data D ₁ (step S77), to calculate the value of virtual data D ₀₁ by using the data D ₀ and the data D _1, and Mod_x (step S78). Then, determination of mod_y is performed (step S79).

Here, if mod_y is 0 (S79)
There Y), the desired dot is on the horizontal line of data D ₀ and the data D _1. That is, in this case, the data _D01 is returned as desired dot data. Also, mod_y
There Otherwise 0 (S79 is N), the data D ₂ and the data D
₃ is read (steps S81 and S82) and data D
₂ and data D ₃ and mod_y to use virtual data D ₂₃
Is calculated (step S83). And this data D
₂₃ , data D ₀₁₂₃ is obtained from the previously obtained data D ₀₁ and mod_y, and this is returned as the value of the desired dot.

As a result, unnecessary interpolation calculation for a desired dot which does not require interpolation calculation is eliminated, so that interpolation processing can be performed efficiently. As a result, the amount of calculation is reduced, and the processing speed is increased.

In the above example, the interpolation calculation at the time of image size conversion and the like has been described as an example. However, for example, even when RGB data is converted to YMC data, one point in the space is converted to a known point in the vicinity. Since the interpolation calculation of obtaining from the point is required, the above processing may be applied to such a case. Also in this case, the speed of the data conversion process can be increased.

[0079]

As described above, according to the present invention,
Since the work memory is set to 2 ⁿ lines, a processing flow that would normally be complicated by reducing the work memory can be configured extremely easily, and therefore, an image processing apparatus with a small memory capacity and high-speed size conversion processing is provided. It becomes possible. In addition, since the coordinates of the interpolation calculation can be easily calculated even if a memory smaller than one frame is used, the interpolation calculation can be performed in parallel with the expansion calculation when the image compressed by JPEG or the like is expanded. Therefore, the size conversion process can be performed at a higher speed. In addition, since the work memory can be arbitrarily reduced under the setting conditions of 2 ⁿ lines, for example, when the data size of the source image is 640 × 480 dots, it is 96% as compared with the conventional case.
, And therefore the degree of saving of the memory capacity is extremely large, which can contribute to cost reduction. In addition, since the interpolation processing is performed while discriminating the positions of the dots that do not require the calculation, the calculation processing frequently used in the image size conversion can be efficiently performed, thereby reducing the amount of calculation as a whole. Thus, higher-speed image processing can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of an image processing apparatus according to a first embodiment.

FIG. 2 shows yellow (magenta, cyan) used for data expansion and size conversion used in the first embodiment.
FIG. 3 is a diagram showing an address map of a work memory area (YMC work area) for use.

FIG. 3 is a diagram showing a relationship between an address of real data and a real coordinate to be developed in a YMC work area and a state of development of the real data.

FIG. 4 is a diagram (part 1) illustrating a relationship between real data and virtual coordinates sequentially developed for every eight lines in a YMC work area and a developed state of the real data.

FIG. 5 is a diagram (part 2) illustrating a relationship between real data and virtual coordinates sequentially developed every eight lines in a YMC work area and a developed state of the real data.

FIG. 6 is a diagram (part 3) illustrating a relationship between real data and virtual coordinates sequentially developed every eight lines in a YMC work area and a developed state of the real data.

FIG. 7 is a diagram in which the real data development diagram of FIG. 3 is re-shown in virtual coordinates.

FIG. 8 is a flowchart illustrating an operation of a process according to the first embodiment.

9 is a flowchart (part 1) for explaining the processing of the flowchart in FIG. 8 in more detail;

FIG. 10 is a flowchart (part 2) for explaining the processing of the flowchart in FIG. 8 in more detail;

FIG. 11 is a diagram for explaining that an image having the same size as the source image can be obtained by enlarging the lower right quarter of the source image by interpolation.

FIG. 12 is a flowchart (subroutine) of an interpolation calculation process according to the second embodiment.

FIG. 13 is a diagram showing a relationship between input and output of a subroutine.

FIG. 14 is a diagram illustrating a positional relationship of dots (actual data) for describing processing according to the second embodiment.

FIG. 15 is a diagram schematically illustrating a frame memory area of, for example, 640 × 480 dots in which image data is expanded when performing image processing.

FIG. 16 is a diagram showing a coordinate system newly introduced to represent the position of new data to be obtained by conventional interpolation as one point on the coordinate system.

FIG. 17 is a flowchart illustrating an image processing operation in a conventional printer.

[Explanation of symbols]

 Reference Signs List 1 image processing device 2 print control circuit 3 digital camera (camera) 4 CPU 5 print start key 6 data memory 7 work memory 8 print memory 9 ROM 11 YMC work area A 12 YMC work area B

Claims (8)

[Claims] 1. The number of lines of a memory used for image processing is 2
ⁿ (n = 1, 2, 3,...). 2. The image processing apparatus according to claim 1, wherein the memory has address values represented by continuous numerical values. 3. The coordinates (x, y) of each pixel on a memory used when performing image processing by treating the original image data as a two-dimensional plane are represented by continuous coordinate values (m, n). When the number of lines is set to 2 ⁿ (n = 1, 2, 3,...), The lower n bits of coordinates are used when calculating an address on the memory. . 4. The image processing apparatus according to claim 3, wherein the image processing is an interpolation calculation for converting the size of image data. 5. The image processing apparatus according to claim 3, wherein the interpolation calculation is performed while determining whether or not the interpolation calculation is possible by a predetermined conditional expression. 6. The conditional expression is defined as follows: when the virtual coordinates have a resolution r times the real coordinates, and the number of times of real data development for each half of the number of memory lines is i, If the virtual coordinate value Y is “Y <(2 ⁿ⁻¹ × (i + 1) −
1) × r ”or the actual coordinate value y of the dot of the actual data at the upper left of the target dot is“ y <(i × 2 ⁿ⁻¹ +2 ⁿ⁻¹⁾
The image processing apparatus according to claim 3, wherein it is determined that interpolation calculation is possible when "-1)" is satisfied. 7. The image processing apparatus according to claim 3, wherein the determination based on the conditional expression is not performed at the time of final processing. 8. In an interpolation process for obtaining an arbitrary point in space from known points around the space, it is determined whether or not calculation to be performed in the processing can be omitted. An image processing apparatus which controls so as not to perform the processing.