JPH05284338A - Picture data interpolation method and picture data interpolation device - Google Patents

Abstract

PURPOSE:To obtain an interpolation method in which an edge part is kept smooth and deterioration in the resolution and an extended noise level due to magnification are suppressed by obtaining picture data of a 1st pseudo picture element based on a monotonously increasing nonlinear function. CONSTITUTION:Brightness information of an original picture inputted externally is converted into a digital signal by an external interface 1 and stored in a picture memory 2 as a basic picture element. Then 1st and 2nd function tables are generated by a microcomputer 3 and stored in memories 4,5. Then picture element levels of a 1st pseudo picture element adjacent obliquely to plural basic picture elements forming an original picture are obtained based on a nonlinear function and picture element levels of a 2nd pseudo picture element adjacent longitudinally and laterally to the plural basic picture elements are obtained based on a nonlinear function to obtain a 2nd magnified picture. Through the constitution above, the edge part is kept smooth and interpolation to suppress deterioration in the resolution and an extended noise level due to magnification is implemented.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a television which handles gradation images,
The present invention relates to an interpolation method and apparatus for expanding the number of pixels such as an image in an information device such as a video, a movie, and a printer.

[0002]

2. Description of the Related Art In recent years, with the digitization and high density of video equipment, the importance of image interpolation technology is increasing in order to effectively use existing sources. Area ratio is about 4
In the pixel interpolation technique of doubling, conventionally, a method called linear interpolation is used to calculate an internal division value of 4 or 2 adjacent basic pixels with respect to basic pixels forming an initial image.

The principle of a conventional linear interpolation method will be described below with reference to the drawings. FIG. 27A shows an initial image before interpolation, and ◯ is a basic pixel forming this image. FIG. 27B is an interpolated image, and
Pixels of ▽ and × are interpolated pixels, and are interpolated based on the basic pixel ◯. △ is the average of two vertically adjacent basic pixels, ▽ is the average of two horizontally adjacent basic pixels, and × is the average of four adjacent basic pixels or the average of Δ or the average of ▽. Ask.

[0004]

However, in the above prior art, since the surrounding information is evenly distributed to generate new pixel data, the interpolated band of pixels is different from the band of basic pixels. Deterioration (level difference between adjacent pixels is small), the whole image looks blurry,
There was a problem that the resolution deteriorated.

Further, if there is noise in the basic pixel, the noise is evenly distributed around the basic pixel, so that there is a problem that the noise is expanded. Further, if the diagonal line caused by the density difference in the initial image is enlarged to the extent that the pixel can be seen even if it is a diagonal line when seen by a person, for example, FIG. 28A (each grid is a basic pixel, 40, The numerical value of 80 appears as shown in FIG. 28B. However, in the image subjected to the linear interpolation, it becomes as shown in FIG. 28B, and there is a problem that it looks jagged even if seen by human eyes. ing.

In view of the above problems, it is an object of the present invention to provide an image data interpolating method and apparatus which keeps an edge portion smooth and suppresses deterioration of resolution due to enlargement and enlargement of noise.

[0007]

In order to solve the above problems, the image data interpolation method of the present invention uses the original image composed of a plurality of basic pixels as the first image, and diagonally adjoins the original image. A pixel to be interpolated between basic pixels is a first pseudo pixel, and a pixel to be interpolated between vertically and horizontally adjacent basic pixels is a second pseudo pixel. An image data interpolation method for arranging a second pseudo pixel to form a second image, comprising: a first step of obtaining a pixel level of a first pseudo pixel based on a monotonically increasing nonlinear function; Second procedure for obtaining the pixel level of the second pseudo pixel based on the non-linear function.

An original image composed of a plurality of basic pixels is defined as a first image, and a pixel to be interpolated between the diagonally adjacent basic pixels is a first pseudo pixel, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method for forming a second image by arranging a first pseudo pixel and a second pseudo pixel in the first image, with a pixel to be interpolated as a second pseudo pixel. It may consist of a first procedure for obtaining the pixel level of the first pseudo pixel based on a monotonically increasing nonlinear function and a second procedure for obtaining the pixel level of the second pseudo pixel based on a linear function. ..

In the nonlinear function, the domain of the input value and the range of the mapping value match, the input value and the mapping value match at one point within the range, and the first derivative is convex upward. May be. In the non-linear function, when the input value is the center value of the domain, the mapping value may be a value near the center of the range.
In the first procedure, the first step 1.1 of obtaining the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels diagonally adjacent to the first pseudo pixel, and the pixel level of the four basic pixels Or the average of the pixel levels of two pixels obtained by removing the maximum value max1 and the minimum value min1 from the pixel levels of the four basic pixels.
Using the non-linear function monotonically increasing in the interval [min1, max1] and the second step of calculating as, and the mapping values corresponding to the maximum value max1 and the minimum value min1 are max1 and min1, respectively.
The method comprises: a first 1.3 step of obtaining a mapping value D1 of the non-linear function with respect to an average value ave1; and a 1.4th step of storing the mapping value D1 as pixel data of the first pseudo pixel in an image memory. The procedure of is the maximum value max2 and the minimum value of the pixel level of four pixels adjacent to the second pseudo pixel, that is, two basic pixels and two first pseudo pixels.
Either the 2.1st step of obtaining min2 and the average of the pixel levels of the four pixels, or the average of the pixel levels of two pixels obtained by removing the maximum value max2 and the minimum value min2 from the pixel levels of the four pixels The second step in which one is calculated as the average value ave2, and the mapping values corresponding to the maximum value max2 and the minimum value min2 are max2 and min2, respectively, and the interval [min
2, max2], using a monotonically increasing nonlinear function, the average value ave
The method may include a second step 2.3 for obtaining the mapping value D2 of the nonlinear function with respect to 2 and a second step 2.4 for storing the mapping value D2 in the image memory as the image data of the second pseudo pixel.

An original image composed of a plurality of basic pixels is defined as a first image, and a pixel to be interpolated between the diagonally adjacent basic pixels is a first pseudo pixel, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method for forming a second image by arranging a first pseudo pixel and a second pseudo pixel in the first image, with a pixel to be interpolated as a second pseudo pixel. The pixel level of the first pseudo pixel is determined by the first
1.1th step of obtaining the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels that are diagonally adjacent to the pseudo-pixels, and the average of the pixel levels of the four basic pixels, or the four basic pixels From the pixel level, the maximum value max1
And the second step of calculating one of the averages of the pixel levels of the two pixels excluding the minimum value min1 as the average value ave1, and the difference between the maximum value max1 and the minimum value min1 and the difference between them. Is smaller than a predetermined value, the third step of determining that there is a correlation, and when there is no correlation, the average value a
If ve1 is the pixel level of the first pseudo pixel and there is a correlation, it may also include the step 1.4 for obtaining the pixel level of the first pseudo pixel based on a monotonically increasing nonlinear function.

The determination of the pixel level of the second pseudo pixel is performed by determining the maximum value max2 of the pixel levels of the four basic pixels and the two first pseudo pixels adjacent to the second pseudo pixel. And the second step of obtaining the minimum value min2, and the average of the pixel levels of the four pixels, or the average of the pixel levels of two pixels obtained by removing the maximum value max2 and the minimum value min2 from the pixel levels of the four pixels The second step of calculating one of the two as the average value ave2 and the maximum value m
The second step of calculating the difference between ax2 and the minimum value min2 and determining that there is a correlation when the difference is smaller than a predetermined value, and if there is no correlation, the average value ave2 is set to the second pseudo pixel. And the pixel level of the second pseudo pixel is obtained based on the monotonically increasing non-linear function when the pixel level of the second pixel is correlated.

In the non-linear function, the domain of the input value and the range of the mapping value match, the input value and the mapping value match at one point within the range, and the first derivative is convex upward. .. The image data interpolation method according to claim 10 or 11, characterized in that. 13. The image data interpolation method according to claim 12, wherein when the input value is the center value of the domain, the mapping value of the non-linear function is a value near the center of the range.

The first image is luminance information that forms a color image together with two pieces of color difference information, and the color difference information may be interpolated by an average value of a plurality of adjacent pixels. The image data interpolation method further includes a third step of reducing the image enlarged by the interpolation, and the third step is the step of reducing the length of each pixel row in the first direction. And a third step of reducing the length of each pixel column in the second direction orthogonal to the first direction.

Further, the image data interpolating apparatus of the present invention uses the original image composed of a plurality of basic pixels as the first image, and the first pixel to be interpolated between the diagonally adjacent basic pixels. Pseudo pixel, a pixel to be interpolated between vertically and horizontally adjacent basic pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. An image data interpolating device for forming
An image memory for storing an image and a second image;
Either the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels adjacent to the pseudo pixel, and the average of the four basic pixels, or the average of two pixels excluding the maximum value max1 and the minimum value min1. And a first average / maximum / minimum determining means for calculating the average value ave1 and a monotonically increasing non-linear function in which the minimum and maximum values of the domain and range are both Min1 and Max1 for the average value ave1. First function mapping means for storing a mapping value D1 calculated by the following equation using the first function F1 in the image memory as image data of the first pseudo pixel,

[0015]

[Equation 9]

The maximum value max2 and the minimum value min2 of the pixel levels of the four basic pixels adjacent to the second pseudo pixel and the two first pseudo pixels, and the average or maximum value of the four pixels. Second average / maximum / minimum determining means for obtaining an average value ave2 which is an average of two pixels excluding max2 and minimum value min2, and the average value ave2 are both minimum and maximum values in the domain and range Min2 And a second function mapping means for storing a mapping value D2 calculated by the following equation using a second function F2 which is a monotonically increasing non-linear function Max2 in the image memory as image data of the second pseudo pixel. When

[0017]

[Equation 10]

Is provided.

[0019]

According to the above means, the present invention firstly provides a first average / maximum / minimum determining means (for the pixel levels of the four basic pixels adjacent to the first pseudo pixel read from the image memory). Step) is maximum value max1, minimum value
Find min1 and average value ave1. For the average value ave1,
The first function mapping means (step) is calculated by (Equation 9) using the first function F1 which is a monotonically increasing non-linear function whose minimum and maximum values in the domain and range are both Min1 and Max1. Then, a mapping value D1 is calculated. The mapping value D1 is stored in the image memory as image data of the first pseudo pixel.

Next, regarding the pixel levels of the four basic pixels and the two first pseudo pixels adjacent to the second pseudo pixel read from the image memory, the second average.
The maximum / minimum determining means (step) determines the average value ave2 for the average value ave2, and the second function mapping means (step) determines that the minimum value and the maximum value of the domain are both Min2 and Min2.
The second function F2, which is a monotonically increasing nonlinear function that is Max2
The mapping value D2 calculated by (Equation 10) using is calculated. The mapping value D2 is stored in the image memory as image data of the second pseudo pixel.

[0021]

(First Embodiment) An image data interpolation method and apparatus according to the first embodiment of the present invention will be described below with reference to the drawings. In this embodiment, the case where the input image is a monochrome image having luminance information will be taken up.

FIG. 1 shows the arrangement of an apparatus for executing the image data interpolation method in this embodiment. Reference numeral 1 denotes an external interface that converts the input image data of the first image into an 8-bit digital signal and controls input / output with the outside. An image memory 2 stores the original first image and the enlarged second image. The first image is composed of basic pixels, and the second image is composed of basic pixels and pseudo pixels. The pseudo pixel is a pixel added by interpolating the first image. FIG. 2 shows the positional relationship between the basic pixel and the pseudo pixel. The image memory 2 initially stores only the basic pixel a,
As the process progresses, mapping values D1, D2, D3 described later are sequentially stored as the first pseudo pixel b and the second pseudo pixels c1, c2. (X, y) in FIG. 2 represents the position of the pixel in the image, and this position is indicated by the vertical and horizontal addresses represented by n and m in the image memory 2.

3 is a microcomputer, CP
U, ROM, RAM, and an input / output unit, first average value / maximum value / minimum value determining means (step), first function mapping means (step), second average value / maximum value / minimum It has respective functions of a value determining means (step) and a second function mapping means (step), and performs the interpolation processing of the present invention on the first image.

Reference numeral 4 is a first function table memory, which stores a table of the first function. Reference numeral 5 denotes a second function table memory, which stores a second function table.
FIG. 2 is a schematic flow chart showing the processing performed by the configuration in FIG. 1 for each function. Reference numeral 202 denotes a first average value / maximum value / minimum value determination step, which includes four basic pixels a (x, y), a (x + 1, y), and a (x that are adjacent to the first pseudo pixel b. , y +
1) For a (x + 1, y + 1), the maximum value max1, the minimum value min1, and the average value ave1 of those pixel levels are obtained. Average value ave1
May be an average of two pixels obtained by removing the maximum value max1 and the minimum value min1 from the four basic pixels.

Reference numeral 203 denotes a first function mapping step,
The mapping values corresponding to the maximum value max1 and the minimum value min1 are max1 and
The mapping value D1 of the first function for normalization and the average value ave2 is calculated by the formula shown in (Equation 11), which is a nonlinear function that is min1 and monotonically increases in the interval [min1, max1]. D1 is written in a predetermined position of the image memory 2 as the first pseudo pixel b (x, y).

[0026]

[Equation 11]

Reference numeral 204 denotes a second average value / maximum value / minimum value determination step, which is adjacent to the second pseudo pixel c1 and is 2
For one basic pixel a (x, y), a (x + 1, y) and the two first pseudo-pixels b (x, y), b (x, y-1), the maximum of their pixel levels The value max2, the minimum value min2, and the average value ave2 are calculated.
The average value ave2 may be an average of image data of 2 pixels obtained by removing the maximum value max2 and the minimum value min2 from the image data of the four basic pixels. Also, the second pseudo pixel c2
Similarly, the maximum value max3, the minimum value min3, and the average value av
Find e3.

205 is a second function mapping step,
The mapping values corresponding to the maximum value max2 and the minimum value min2 are max2 and
The mapping value D2 of the second function with respect to the average value ave2 is normalized by the formula shown in (Equation 12) which is a nonlinear function that is min2 and monotonically increases in the interval [min2, max2]. D2 is written in a predetermined position of the image memory 2 as the second pseudo pixel c1 (x, y). Also, for the second pseudo pixel c2 (x, y), the mapping value D3 is similarly obtained by the equation shown in (Equation 13).

[0029]

[Equation 12]

[0030]

[Equation 13]

Regarding the image data interpolating method and apparatus according to the first embodiment of the present invention configured as described above,
The operation will be described with reference to the drawings. The brightness information of the first image input from the outside, which shows the main flow of the interpolation process in FIG. 4, is converted into an 8-bit digital signal by the external interface 1 in FIG. 1 and is used as a basic pixel in FIG. Are stored in even-numbered cells, and the coordinates x and y are initialized (step 30 in FIG. 4).
1). Next, the microcomputer 3 creates a table of the first and second functions. That is, in this embodiment, the functions F1 and F2 are directly used (Equation 11), (Equation 1)
2) and (Equation 13) are not calculated, and the values read from the table are used as the mapping values with the input values of the respective functions as addresses, using a lookup table showing the relationship between the input values of each function and the mapping values. By doing so, each formula is calculated (step 302). FIG. 5 shows an example of input / output characteristics of the tables of the first and second functions in this embodiment.
Both of these two functions are monotonically increasing nonlinear functions,
The mapping values for the maximum value 1.0, the minimum value 0, and the central value 0.5 of the input values match 1.0, 0, and 0.5, respectively, and the first derivative thereof is convex upward.

As described above, in this embodiment, the first and second
The function table of is created by the microcomputer every time an image is input so that various functions can be used, but these function tables can be created in the ROM in advance. If the ROM table is referred to, the table does not have to be created each time an image is captured, step 302 can be omitted, and interpolation processing can be realized at a higher speed.

Thereafter, the first pseudo pixel b (x, y) is obtained (step 303). FIG. 6 shows a detailed flowchart of step 303. First, the first pseudo pixel b
To obtain (x, y), four basic pixels a (x, y) and a (x +
1, y), a (x, y + 1) and a (x + 1, y + 1) are read from the image memory 2 (step 401). Next, the average value ave1, the maximum value max1, and the minimum value min1 of the pixel levels (luminance information) of the four pixels are obtained (step 402). The average value, the maximum value, and the minimum value are normalized so that the value is [admin1, admax1] within the range of the address of the ROM table according to the formula shown in the address ad1 (Equation 14) of the first function table. After that, the first function table is drawn and the table value e1 is obtained (step 403). However, (number 1
In the formula 4), admax1 and admin1 are the maximum and minimum values of the address of the first function table.

[0034]

[Equation 14]

The range of table values is [admin1, admax
1], the table value e is calculated in step 404.
1 is converted to the actual pixel level D1 by the formula shown in (Equation 15) (step 404), and D1 is written in a predetermined position of the image memory 2 (step 405).

[0036]

[Equation 15]

Further, the second pseudo pixel c1 (x, y) is obtained (step 304 in FIG. 4). Step 3 in FIG.
4 shows a detailed flowchart of 04. First, in order to obtain the second pseudo pixel c1 (x, y), two basic pixels a (x, y),
a (x + 1, y) and two first pseudo pixels b (x, y-1), b (x,
y) is read from the image memory 2 (step 501 in FIG. 7). Next, the average value ave of the brightness information of these 4 pixels
2, the maximum value max2 and the minimum value min2 are obtained (step 5
02). From the average value, the maximum value, and the minimum value, from the formula shown in the address ad2 (Equation 16) of the second function table, RO
After normalization so that the value becomes [admin2, admax2] within the range of the address of the M table, the first function table is drawn and the table value e1 is obtained (step 503).
In the equation (16), admax2 and admin2 are the maximum and minimum addresses of the second function table.

[0038]

[Equation 16]

The range of table values is [admin2, admax
2], the table value e2 is converted to the actual pixel level D2 by the formula shown in (Equation 17) (step 504),
D2 is written in a predetermined position of the image memory 2 (step 505).

[0040]

[Equation 17]

Subsequently, the second pseudo pixel c2 (x, y) is obtained (step 305 in FIG. 6). Step 3 in FIG.
The detailed flowchart of 05 is shown. First, in order to obtain the second pseudo pixel c2 (x, y), two basic pixels a (x, y),
a (x, y + 1) and two pseudo pixels b (x-1, y) and b (x, y) are read from the image memory 2 (step 506). Next, as in step 304 above, the average value ave3 and the maximum value
max3, minimum value min3, the address ad3 of the second function table from the expression shown in (Equation 18), the table value e3 from the second function table, and the pixel level D3 from the expression shown in (Equation 19)
Is calculated (steps 507 to 509).

[0042]

[Equation 18]

[0043]

[Formula 19]

The steps 303 to 305 are performed in the x direction by the steps 306 and 307, and the steps 308 and 309.
Thus, in the y direction, interpolation is sequentially performed until the predetermined number of pixels is completed. Note that the pixels at the edge of the image (for example,
From (x, y) = (0,0)) to the second pseudo pixel (c1 (0,0),
This is because the first pseudo pixel (b (0, -1), b (-1,0)) that does not exist in the image memory 2 is required when the image data of c2 (0,0)) is created. In this case, the image data of the second pseudo pixel is the average value of the image data of the basic pixels (for c1 (0,0), the average of a (0,0) and a (1,0), c2 (0, A) for 0)
The average of (0,0) and a (0,1) is used.

Further, (Equation 20) shows the non-linear function used for the first and second functions of this embodiment. However, the constants are a = 0.98, b = 0.01, and c = 0.22.

[0046]

[Equation 20]

Next, the specific effect of this embodiment will be described. First, the effect of reducing resolution deterioration and the smoothness of diagonal lines will be described. First, FIG. 9 shows the results of interpolation by this method and the conventional linear method when luminance information having different levels is input in the oblique direction as shown in FIG. Figure 9
(A) is a linear system and (b) is a system.
In both the linear method and the present method, bold numbers are basic pixels and other numbers are pixel levels obtained by interpolation, and average values of pixel levels are also shown for each diagonal direction 1 to 5. Note that here, interpolation is performed assuming that this line is continuous even outside the range of FIG.

Based on the figure, in order to compare the change in the level of the boundary portion by interpolation, in the basic image, the boundary (the one-dot chain line in the direction 3) and two lines before and after the boundary (direction 1,
The result of obtaining the average of the pixel levels on the two-dot chain line (2, 4, 5) is shown in FIG. In FIG. 10, ∘ indicates the average of pixels in each direction according to this method, and × indicates the average in the case of linear. From this figure, compared to the case where the present method is based on the linear method, the edges are raised and the boundaries are more preserved. That is, it can be seen that the resolution of the basic image is maintained more.

FIG. 11 shows how a person looks at this boundary. The thick solid line in the figure is the boundary between levels 40 and 60. When the linear method is used (Fig. 11 (a)), there is no continuous boundary with a level difference of about 20, and there is a soft look at level 50. That is, the boundaries are recognized as large wavy lines shown in the figure, which causes the diagonal lines to appear jagged.

On the other hand, in the case of this method (see FIG. 11)
(B)) Since the level difference is continuous with about 20, this boundary becomes a fine wavy line as shown in the figure, and the slanted line looks much smoother than the linear line. Next, the principle of noise reduction of this method will be described. When the noise is included in the four basic pixels, the noise is likely to have the maximum value or the minimum value among the four basic pixels because the level is usually different from that of the other three pixels.

At this time, if the luminance of the noise in the four pixels is greatly different from that of the other three pixels by supplementing with the linear average as in the conventional case, the luminance difference is evenly distributed and the second image is large. It becomes noise. On the other hand, in this method, the maximum value and the minimum value are used for normalization, and the average value is used to draw the first or second function table. At that time, as shown in FIG. The first and second functions based on such a non-linear function are such that when the average value is close to the maximum value, the mapping value is mapped to a value closer to the maximum value, and the value close to the minimum value is mapped to a value closer to the minimum value. have.
(In other words, when the average value is far from the maximum value and the minimum value, the mapping value becomes a value farther from the maximum value and the minimum value.) As a result, the luminance of noise in the four pixels is large compared with the other three pixels. If they are different, the mapping value is such that the influence of this noise is almost ignored. In other words, the expansion of noise is greatly suppressed in this method compared to the linear method.

In this embodiment, when a black-and-white image of 480 × 640 dots is enlarged by interpolation and the interpolation method described in this embodiment is executed, smooth diagonal lines which cannot be realized by conventional linear interpolation are realized, Second with good image quality
I was able to get the image. Further, in the present embodiment, the first and second functions are normalized by the maximum value and the minimum value. However, if this is not performed, the noise reduction effect is lost and the resolution may not be improved. Do you get it.

Further, in the present embodiment, both the first function and the second function have the maximum value of 1, the minimum value of 0 and the center of 0.5, the input value and the mapping value match, and the center of the domain is 0. In the range smaller than 5, the function is convex downward, and in the range larger than the center of the domain, the function is convex upward, and a monotonically increasing function is used. The present invention is not limited to this, and it is also possible to use a function in which the input value and the mapping value match at three points other than the maximum value 1, the minimum value 0 and the center of the domain as each function. At this time, the degree of brightness of the second image can be changed. However, in this case, it is desired that the input value is 0.4 or more and 0.6 or less so as to match the mapping value. This is because if the image is mapped outside this range, the luminance of the basic pixel and the pseudo pixel greatly differ, the image looks like a mosaic, and the image quality greatly deteriorates.

Further, with respect to the image data of the basic pixel shown in FIG. 8, the first pseudo pixel is calculated by the mapping of the first function shown in (Equation 20), and the second pseudo pixel is adjacent. Two
When the average value of the two basic pixels is used, it is the same as that shown in this method of FIG. 9, and the image quality can be improved, and the second pseudo pixel can be calculated only by averaging the two pixels. It takes less time and can be calculated faster than the above.

FIG. 1 shows the result when another interpolation process is performed on the basic pixel image data shown in FIG.
2, shown in FIG. In FIG. 12A, the first pseudo pixel is calculated by the mapping of the first function shown in (Equation 12), and the second pseudo pixel is calculated.
12 shows the case where the pseudo pixel of FIG. 12 is an average value of 4 pixels of two adjacent basic pixels and two first pseudo pixels.
In (b), the first pseudo pixel is the average value of four adjacent basic pixels, and the second pseudo pixel is calculated by the mapping of the second function shown in (Equation 12).

In FIG. 12A, it can be seen that the resolution is retained and the diagonal lines are smoother than the conventional linear shape of FIG. However, in FIG. 12B, the resolution is reduced as in the conventional linear processing, and the slanted line is also jagged. Therefore, only the first pseudo pixel is
It is necessary to use a mapping value with a non-linear function, and that is sufficient.

In FIG. 13, using the first and second functions shown in FIG. 5 described above, the average value given to these non-linear functions is calculated using four basic pixels or two basic pixels and two pseudo pixels. The case where the mapping value by the average value of the remaining two pixels excluding the pixel of the maximum value and the pixel of the minimum value among the image data of the four pixels is used as the luminance information of the first and second pseudo pixels is shown. The resolution of diagonal lines can be saved most and smooth diagonal lines can be obtained.

In this embodiment, the same function is used as the first and second functions, but it is of course possible to use different functions. In this embodiment, as a non-linear function (Equation 20)
Although the function as shown in FIG. 3 is used, any monotonically increasing function in which the maximum value, the minimum value, and the center of each of the domain and the range are 1, 0, and 0.5 may be used. For example, a higher-order function such as a cubic function shown in (Equation 21) or a 3rd order shown in (Equation 22)
A function including an angular function can also be used. In this function of (Equation 20), the constant a is adjusted so that the value of the first derivative y ′ is positive when x is between 0 and 1.

[0059]

[Equation 21]

[0060]

[Equation 22]

Further, in this embodiment, the image data of the basic pixel, the first pseudo pixel and the second pseudo pixel are stored in the image memory in the adjacent memory addresses as in the arrangement on the image. However, it is of course possible to store only the image data of the basic pixel or only the image data of the pseudo pixel adjacently in the memory in the image memory. Further, although the first image and the second image are stored in the same image memory, it is of course possible to adopt a configuration using a plurality of image memories. (Second Embodiment) An image data interpolation method and apparatus according to a second embodiment of the present invention will be described below with reference to the drawings. In this embodiment, the image data is brightness and 2
Take the case of a full-color image with one color difference information.

FIG. 14 shows the structure of the image data interpolating apparatus in this embodiment. This configuration is basically the same as the configuration of the first embodiment shown in FIG. 1, except that the external interface 1101, the image memory 1102, and the microcomputer 1103 are compatible with full-color images. Only different points will be described below.

The external interface 1101 converts the luminance information Y, which is the image data of the first image input from the outside, and the two color difference information RY and BY into 8-bit digital signals. Image memory 1102
Is an image memory for storing the first image and the enlarged second image, and has three pages for storing respective information, and the page 1 (p = 0) contains the luminance information Y and the page 1
In (p = 1), the color difference information RY is stored in page 2 (p = 2).
Stores color difference information BY.

The microcomputer 1103 is a CP
U, ROM, RAM, and an input / output unit, first average / maximum / minimum determining means for luminance information, first function mapping means, second average / maximum / minimum determining means, and second function mapping For the color difference information, it further has an averaging means for obtaining the average of the color difference information of adjacent pixels, and performs the interpolation processing of the present invention on the first image.

Regarding the image data interpolating method and apparatus according to the second embodiment of the present invention configured as described above,
The operation will be described with reference to the drawings. The basic pixel of the luminance and the two color difference information of the first image input from the outside, which shows the main flow of the interpolation processing in FIG. 15, are converted into 8-bit digital signals by the external interface 1101, and each page of the image memory 1102 is converted. Are stored in even addresses of. Along with this, the coordinate x,
Initialization of y is performed (step 120 in FIG. 15).
1). Next, a table of the first and second functions is created by the microcomputer 1103 (step 1
202). The input / output characteristics of this function are the same as in FIG. 5 described above. Then, the page of the image has p =
The image memory 110 is set to 0 (step 1203).
The second page is switched to the page having the value of P at that time (step 1204).

Next, the first pseudo pixel b (x, y) is obtained (step 1205). FIG. 16 shows a detailed flowchart of step 1205. 4 basic pixels a (x, y)
, A (x + 1, y), a (x, y + 1), and a (x + 1, y + 1) image data are read from the image memory 1102 (step 1301 in FIG. 16). Then, it is determined whether these pieces of information are luminance information or color difference information (whether P = 0 or not) (step 13).
02), in the case of luminance information (when P = 0), the average value ave1, the maximum value max1, and the minimum value min1 of these are obtained (step 1303). From the average value, the maximum value, and the minimum value, the address ad1 of the first function table is calculated by the formula shown in (Equation 14), and the table value e1 is obtained by subtracting the first function table (step 1304). ). Table value
The e1 is converted into the brightness level D1 of the actual pixel by the formula shown in (Equation 15) (step 1305). This mapping value D1 is used as the first pseudo pixel b (x, y) in the image memory 11
02 at a predetermined position (step 130)
6).

If the image data is color difference information in step 1302 (if P = 0), the color difference information of the basic pixels is averaged (step 1307), and this average value is calculated as the first pseudo pixel b ( image memory 11 as x, y)
02 at a predetermined position (step 130)
6). Further, the second pseudo pixel c1 (x, y) is obtained (step 1206 in FIG. 15). Step 12 in FIG.
6 shows a detailed flowchart of 06.

Two basic pixels a (x, y), a (x + 1, y) and 2
Image data of four pixels of one first pseudo pixel b (x, y-1) and b (x, y) is read from the image memory 1102 (step 1401 in FIG. 17). It is judged whether the image data of these 4 pixels is luminance information or color difference information (whether P = 0 or not) (step 1402), and when it is luminance information (when P = 0), the average value ave2 and maximum value of these max2 and minimum value min2 are obtained (step 1403), and the address ad2 of the second function table is calculated from the average value, the maximum value, and the minimum value (Equation 1).
The table value e2 is obtained by subtracting the second function table calculated from the formula shown in 6) (step 140).
4). The table value e2 is converted to the brightness level D2 of the actual pixel by the formula shown in (Equation 17) (step 140).
5). This mapping value D2 is written in a predetermined position of the image memory 1102 as the second pseudo pixel c1 (x, y) (step 1406).

If the image data is color difference information in step 1402, the average of the color difference information of four pixels is obtained (step 1407). This average value is written as a second pseudo pixel c1 (x, y) at a predetermined position in the image memory 1102 (step 1406). After this, the second pseudo pixel c2 (x, y) is obtained (step 12 in FIG. 15).
07). FIG. 18 shows a detailed flowchart of step 1207.

Similar to the method for obtaining the second pseudo pixel c1 (x, y) described above, in the case of luminance information, the mapping value D3 is calculated, and this is calculated as the second pseudo pixel c2 (x, y). In the case of color difference information, the average is calculated, and this is calculated as the second pseudo pixel c2 (x, y).
Is written in a predetermined position of the image memory 1102 (steps 1501 to 1506). Second pseudo pixel c
When 2 (x, y) is obtained, the above process is performed on the third page (P =
It is checked whether or not up to 2) (P> 1) (step 1208), and if not completed (P = 0,
1), the next page is set (step 120).
9), the same process as above is repeated for the next page.

Further, steps 1210, 1211, 12
12 and 1213, interpolation is sequentially performed on the pixels at the x and y positions, and finally the image memory 110
All the image data of the second image is formed in 2. Next, the effect of this embodiment will be described. In this example,
The brightness information is processed in the same manner as in the first embodiment. As shown in the effect of the first embodiment, noise is not increased, smooth diagonal lines are reproduced, and deterioration of resolution is small. There is an effect such as.

Since the same processing as in the linear system is performed on the two color difference information R-Y and B-Y, the hue and saturation are not significantly different from the original ones, and good color reproduction can be realized. .. However, since the same processing as the linear method is performed on the color difference information, there is no effect like luminance on the resolution of the color signal. However, since the human visual characteristics with respect to the color resolution are not as good as the luminance resolution, the linear interpolation is sufficient. In addition, since the calculation speed of the color difference information is a linear method, the interpolation processing can be performed at a higher speed.

When the number of pixels is increased by using this interpolation method for a full-color image having the number of pixels of 480 × 640,
A smooth diagonal line was realized, and a good full-color image having about four times the number of pixels could be created. It should be noted that from the image data of the pixel at the edge of the image (for example, when (x, y) = (0,0))
Of the first pseudo pixel (b (0, -1), b (-1) that does not exist in the image memory 2 when the image data of the pseudo pixel (c1 (0,0), c2 (0,0)) of , 0) image data is required, so the second pseudo pixel in this case is a (0,0) and a (0 for the average value (c1 (0,0)) of the image data of the basic pixels. The average of 1,0) and the average of a (0,0) and a (0,1) for c2 (0,0) are used.

In the second embodiment, the color difference information is averaged by the basic pixel information, but this may be used by using the basic pixel information and the pseudo pixel information. Further, in the present embodiment, the input of full-color image information is the luminance information Y and the two color difference information of RY and BY, but the effect on resolution is the same when inputting RGB information or CMY information. Becomes However, it has a drawback that the hue and saturation slightly change as in the case where the interpolation of the present invention is performed on RY and BY. Therefore, the RGB information and the CMY information are first converted to the luminance information Y. Then, it is desirable to convert the color difference information into two pieces of color difference information RY and BY, perform the interpolation processing described in the second embodiment of the present invention, and convert again into RGB information or CMY information.

In the present embodiment, the first function and / or the second function is a non-linear function. However, when a function combining the non-linear function and / or the linear function is used,
For example, the same effect can be obtained with a function as shown in FIG. However, since a pseudo contour is likely to occur in a portion where the derivative of the function becomes discontinuous, it is desirable to use a non-linear function whose first derivative is continuous.

In this embodiment, the first pseudo pixel and the second pseudo pixel
Although one non-linear function is used to obtain each of the pseudo-pixels, a plurality of non-linear functions can be selectively used according to the difference between the minimum value and the maximum value of the input. In the present embodiment, the first pseudo pixel is obtained and then the second pseudo pixel is obtained. However, it is of course possible to obtain the second pseudo pixel after obtaining all the first pseudo pixels.

Further, in the present embodiment, only the interpolation for increasing the number of pixels by about 4 is shown, but the number of pixels can be further increased by repeatedly applying the present interpolation method to the second image. You can (Third Embodiment) The third embodiment of the present invention will be described below.
A description will be given with reference to the drawings. In this embodiment, the case where the input image is a monochrome image having luminance information will be taken up.

The structure of the apparatus for executing the image data interpolation method in this embodiment is almost the same as that of the first embodiment shown in FIG. 1, except that the microcomputer 3 has a reducing means (step) and a reducing means. The difference is that the function of rate indicating means (step) is added. FIG. 20 is a schematic flow chart showing the processing performed by the present embodiment by function.
This drawing is basically the same as the schematic flow chart of the first embodiment shown in FIG. 3, but the reduction ratio instruction step 2106,
The difference is that a reduction step 2107 is added. Only different points will be described below.

In the reduction rate instruction step 2106, the reduction rate k for the image enlarged by the interpolation is set, and the reduction rate is instructed to the reduction means. The reduction rate k is set to an arbitrary value larger than 1/2 and smaller than 1. Reduction step 2
107 reduces the image enlarged by interpolation at the reduction rate k instructed from the reduction instruction step.

The operation of the data interpolation method and apparatus according to the present embodiment configured as described above will be described. It is assumed that the image data enlarged by interpolation in exactly the same way as in the first embodiment is stored in the image memory 2 as the original image data before reduction. 21A and 21B show simple examples of the original image data and the image data after reduction when the reduction ratio k = 3/4. In the figure, m,
n is an address in the horizontal and vertical directions. Reduction step 2
Reference numeral 107 performs pixel reduction processing in the m direction and pixel reduction processing in the n direction on the original image as shown in FIG. A detailed operation flow of this reduction processing is shown in FIG.

First, an arbitrary value is set in advance as the reduction rate k (step 2201), and m of the image address is set.
Prior to the reduction processing in the direction, n = 0 and m = 0 are set as the initial values of the address (m, n) at which the reduction is started in the original image (step 2202). For the pixel at the address (m, n) currently set in the original image, the position of the pixel in the image after reduction corresponding to the address in the m direction is calculated by the following formula (step) 2203). Pm = (1 / k) m When the number of pixels of the original image is 4x4 and the reduction ratio k = 3/4, the corresponding position P of the pixel after reduction in the m direction with respect to the original image
m is shown in FIG. In the figure, ◯ indicates a pixel, and Pm indicates the position of the pixel to be obtained in the original image in the m direction. Since the value of Pm does not always match the existing address (not necessarily an integer), the m-direction addresses m0 and m1 of the existing pixels before and after that position are obtained by the following formula (step 2004). m0 = int (Pm) However, int truncates n after the decimal point. m1 = m0 + 1 Further, the position Pm of the pixel to be obtained and the existing 2 before and after
Pixel (m0, n), the ratio of each distance to (m1, n) is Pm-m0
: 1- (Pm-m0). From this, the pixel data D (m ', n) to be obtained is obtained by the following formula that internally divides the pixel levels of the front and rear two pixels by the inverse ratio of the distance 1- (Pm-m0): Pm-m0. The value is stored in the address (m ', n) of the image memory 2 (step 2005). D (m ', n) = {1- (Pm-m0)} D (m0, n) + (Pm-mo) D (m1, n) = (D (m1, n) -D (m0, n) } (Pm-m0) + D (m0, n) Then, the next pixel in the m direction is set (step 200
6, 2007), and the above steps 2003 to 2005 reduce one line in the m direction. When the reduction of one line in the m direction is completed, 1 is added to n (step 200
8, 2009), the next one row in the m direction is reduced by one row in steps 2002 to 2007, and n
Perform m-direction reduction for all rows. Figure 23
An example of an image reduced in the m direction is shown in (b). In the figure, ∇ indicates a pixel after reduction, and ∘ indicates a pixel corresponding to the original image.

When the reduction in the m direction is completed, the reduction is also performed in the n direction in the same manner as described above (steps 2010 to 20).
16). FIG. 24A shows an example of image data before being reduced in the n direction, and FIG. 24B shows an example of image data after being reduced in the n direction. ∇ indicates a pixel before reduction, and □ indicates a pixel after reduction in the n direction. Thus, the reduction ratio 3/4 shown in FIG.
The image data reduced by is obtained.

A reduced image P3, image data before reduction (image data enlarged by interpolation) P2,
Explaining the relationship with the image data P1 before being interpolated,
If the expansion rate from P1 to P2 is 2 times and the reduction rate from P2 to P3 is 3/4 times, 2 × 3/4 = 2/3 times from p1 to P3. Note that the reduction ratio k is set to a value larger than 1/2 and smaller than 1. When the magnification is 1 or more, it is better to enlarge the image again by the method described in the first embodiment and reduce the image. This is for two reasons: it is advantageous, and when the magnification is smaller than ½, it is advantageous in terms of resolution and processing time to reduce the original image before enlargement.

Next, the specific effects of this embodiment will be described. The first embodiment has a drawback in that it can only be magnified by a power of 2 in both vertical and horizontal directions, but this embodiment solves this, and the original image can be magnified at any magnification. .. That is, from P1 to P2
An appropriate enlargement ratio can be set by appropriately selecting the combination of the enlargement ratio to P2 and the reduction ratio from P2 to P3.

Although only the black and white image has been described in the present embodiment, it is possible to easily enlarge the full-color image by an arbitrary magnification by the same process. (Fourth Embodiment) A fourth embodiment of the present invention will be described below with reference to the drawings. In this embodiment, the case where the input image is a monochrome image composed of luminance information will be taken up. The structure of the apparatus for executing the image data interpolation method in this embodiment is almost the same as the structure of the first embodiment shown in FIG.

FIG. 25 is a flow chart showing the processing performed by this embodiment for each function. This is basically the same as the first embodiment shown in FIG. 3, but the correlation determining means (step) 3106 and the pseudo pixel selecting means (step) 3107.
Is added, and only different points will be described. In the correlation detection step 3106, if the difference between the maximum value and the minimum value is smaller than the predetermined value, it is determined that there is a correlation.

In the pseudo pixel selection step 3107, the average value is output as a pseudo pixel if it is determined that there is a correlation, and the mapping value is output as a pseudo pixel if it is determined that there is no correlation. The operation of the data interpolation method and apparatus according to the present embodiment configured as above will be described. First in FIG.
The processing flow for obtaining the pseudo pixel b (x, y) of is shown. The same applies when other pseudo pixels are obtained.

The average value, the maximum value, and the minimum value of the four basic pixels are calculated (steps 3201 and 320).
2) Further, the difference between the maximum value and the minimum value is obtained (step 3203). Based on this difference, it is determined whether or not there is a correlation between the pixel to be interpolated and the original pixel. That is, if this difference is smaller than a predetermined value, it is determined that there is correlation, and if it is large, it is determined that there is no correlation (step 3204).

If it is determined that there is a correlation, the average value obtained in step 3202 is written in the image memory 2 as the level of the pixel to be interpolated (step 3).
207). If it is determined that there is no correlation, the first function table is referred to based on the average value (step 32
05), the pixel level is calculated (step 3206),
It is written in the image memory 2 (step 3207).

Next, the specific effects of this embodiment will be described. In this embodiment, since the average value is used as it is as a pseudo pixel in a portion where there is almost no pixel level difference in the image, the processing time can be considerably shortened as compared with the case where the mapping value is used. In addition, since the deterioration of resolution in a portion having almost no level difference cannot be recognized by human eyes, interpolation can be processed at high speed without deterioration of image quality even in the present embodiment.

[0091]

As described above, according to the present invention, by obtaining the image data of the first pseudo pixel by using the non-linear function, the edge portion is kept smooth so that the oblique line does not look jagged, and the enlargement is performed. There is an effect that it is possible to perform the interpolation that suppresses the deterioration of the resolution due to and suppresses the noise expansion. At this time, regarding the calculation of the second pseudo pixel,
A linear average of adjacent pixels or a first non-linear function that is the same as or different from that used to calculate the first pseudo pixel
The same procedure as for the pseudo pixel can be performed.

In particular, for a color image, interpolation using a non-linear function is performed for the luminance information of the pseudo pixels, and the average value of the color difference information of each adjacent pixel is used for the color difference information of the pseudo pixels. The effect is that a good color image with no change in saturation can be created. Further, by further reducing the image once enlarged, it is possible to enlarge the image to an arbitrary size with little deterioration in resolution.

Further, by switching the interpolation by the non-linear function and the interpolation by the average value according to the correlation of the pixels to obtain the pseudo pixel, it is possible to perform high-speed processing without deterioration of resolution.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an image data interpolation device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a positional relationship between a basic pixel a, a first pseudo pixel b, and second pseudo pixels c1 and c2.

FIG. 3 is a flowchart showing a procedure of interpolation processing in the first embodiment of the present invention.

FIG. 4 is a main flowchart of an image data interpolation method in the same embodiment.

FIG. 5 is a diagram of input / output characteristics of first and second functions in the first and second embodiments.

FIG. 6 is a flowchart for obtaining a first pseudo pixel b in the first embodiment.

FIG. 7 (a) is a second pseudo pixel c1 in the same embodiment.
Flowchart to ask. (B) A flowchart for obtaining the second pseudo pixel c2 in the embodiment.

FIG. 8 is a diagram showing an example of basic pixels for obtaining an interpolation characteristic in the embodiment.

9A is a diagram of an interpolation result for the example of FIG. 6 by a linear method. (B) A diagram of an interpolation result for the example of FIG. 6 according to the embodiment.

FIG. 10 is a characteristic diagram of a boundary portion in the same embodiment and the linear interpolation method.

FIG. 11A is a diagram showing a visual appearance of an oblique boundary according to a linear method. (B) A view of the visual appearance of a diagonal boundary according to the embodiment.

FIG. 12A is a diagram when the first function is non-linear and the second function is linear (four-point average) in the embodiment.
Of the interpolation result for the example of FIG. (B) In the same Example, the figure of the interpolation result with respect to the example of FIG. 6 when making a 1st function linear (4 point average) and making a 2nd function nonlinear.

FIG. 13 is a diagram of an interpolation result for the example of FIG. 6 when the mapping value is obtained by averaging the other two pixels of the four pixels excluding the maximum value and the minimum value in the embodiment.

FIG. 14 is a configuration diagram of an interpolation recording device according to a second embodiment of the present invention.

FIG. 15 is a main flowchart of an image data interpolation method in the embodiment.

FIG. 16 is a flowchart for obtaining a first pseudo pixel b in the same example.

FIG. 17 is a flowchart for obtaining a second pseudo pixel c1 in the example.

FIG. 18 is a flowchart for obtaining a second pseudo pixel c2 in the example.

FIG. 19 is a diagram of a function combining a nonlinear function and a linear function.

FIG. 20 is a main flowchart of an image data interpolation method according to the third embodiment of the present invention.

FIG. 21 is a diagram showing a simple example of original image data and image data after reduction in the case where the reduction rate k = 3/4 in the embodiment.

FIG. 22 is a flowchart for reducing an image according to the embodiment.

FIG. 23 is a case where the number of pixels of the original image is 4 × 4 in the same example.
FIG. 7 is a diagram showing corresponding positions Pm of pixels after reduction in the m direction with respect to the original image when the reduction ratio is k = 3/4.

FIG. 24 is a diagram illustrating that the number of pixels of the original image is 4 × 4 in the embodiment.
FIG. 7 is a diagram showing corresponding positions Pn of pixels after reduction in the n direction with respect to the original image when the reduction ratio is k = 3/4.

FIG. 25 is a main flowchart of an image data interpolation method according to the fourth embodiment of the present invention.

FIG. 26 is a flowchart for obtaining a pseudo pixel by performing correlation detection and selection in the embodiment.

FIG. 27 is a principle diagram of a conventional linear interpolation method.

FIG. 28 (a) is a diagram showing a visual appearance of a diagonal boundary in a basic image. (B) The figure of the conventional appearance of the interpolation result about (a) and the visual appearance of the diagonal boundary.

[Explanation of symbols]

 1 External Interface 2 Image Memory 3 Microcomputer 4 First Function Table Memory 5 Second Function Table Memory 1101 External Interface 1102 Image Memory 1103 Microcomputer 1104 First Function Table Memory 1105 Second Function Table Memory

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification code Internal reference number FI technical display location G09G 5/36 9177-5G

Claims (29)

[Claims] 1. An original image composed of a plurality of basic pixels is used as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method in which a pixel to be interpolated between pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. From the first procedure of obtaining the pixel level of the first pseudo pixel based on the monotonically increasing nonlinear function, and the second procedure of obtaining the pixel level of the second pseudo pixel based on the monotonically increasing nonlinear function. An image data interpolation method characterized by the following. 2. An original image composed of a plurality of basic pixels is defined as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method in which a pixel to be interpolated between pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. A first procedure for obtaining the pixel level of the first pseudo pixel based on the monotonically increasing non-linear function, and a second procedure for obtaining the pixel level of the second pseudo pixel based on the linear function. Characteristic image data interpolation method. 3. The non-linear function has a domain of input values and a domain of mapping values that match, the input value and its mapping value match at one point within the range, and the first derivative is convex upward. The image data interpolation method according to claim 1 or 2, wherein 4. The image data interpolation method according to claim 3, wherein when the input value is the center value of the domain, the mapping value of the non-linear function is a value near the center of the range. 5. The first procedure is to obtain a maximum value max1 and a minimum value min1 of pixel levels of four basic pixels diagonally adjacent to the first pseudo pixel.
1 step, and the pixel level average of the four basic pixels, or the four
The 1.2th step of calculating one of the averages of the pixel levels of the two pixels excluding the maximum value max1 and the minimum value min1 from the pixel level of one basic pixel as the average value ave1, and the maximum value max1 and the minimum value min1. The mapping values corresponding to are max1 and
Using a non-linear function that is min1 and monotonically increasing in the interval [min1, max1], the 1.3rd step of obtaining the mapping value D1 of the non-linear function with respect to the average value ave1, and the mapping value D1 is the first pseudo pixel. The second step is to store four basic pixels and two first pseudo pixels that are adjacent to the second pseudo pixel. Maximum pixel level of pixel
The second step of obtaining max2 and the minimum value min2, the average of the pixel levels of the four pixels, or the maximum value max2 and the minimum value min2 from the pixel levels of the four pixels
The second step of calculating either one of the averages of the pixel levels of the two pixels except for the average value ave2, and the mapping values corresponding to the maximum value max2 and the minimum value min2 are max2 and max2, respectively.
The second step of obtaining the mapping value D2 of the non-linear function with respect to the average value ave2 by using the non-linear function which is min2 and monotonically increases in the interval [min2, max2], and the mapping value D2 is set to the second pseudo pixel. 5. The image data interpolation method according to claim 4, further comprising the step of storing the image data as image data in an image memory. 6. An original image composed of a plurality of basic pixels is used as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method in which a pixel to be interpolated between pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. And a second procedure for obtaining the pixel level of the second pseudo pixel and a second procedure for obtaining the pixel level of the second pseudo pixel. 1. Find the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels that are diagonally adjacent to each other.
1 step, and the pixel level average of the four basic pixels, or the four
The 1.2th step of calculating one of the averages of the pixel levels of the two pixels excluding the maximum value max1 and the minimum value min1 from the pixel level of one basic pixel as the average value ave1, and the maximum value max1 and the minimum value min1. The mapping values corresponding to are max1 and
Using a non-linear function that is min1 and monotonically increasing in the interval [min1, max1], the 1.3rd step of obtaining the mapping value D1 of the non-linear function with respect to the average value ave1, and the mapping value D1 is the first pseudo pixel. The second step is to store four basic pixels and two first pseudo pixels that are adjacent to the second pseudo pixel. Maximum pixel level of pixel
The second step of obtaining max2 and the minimum value min2, the average of the pixel levels of the four pixels, or the maximum value max2 and the minimum value min2 from the pixel levels of the four pixels
The second step of calculating either one of the averages of the pixel levels of the two pixels except for the average value ave2, and the mapping values corresponding to the maximum value max2 and the minimum value min2 are max2 and max2, respectively.
The second step of obtaining the mapping value D2 of the non-linear function with respect to the average value ave2 by using the non-linear function which is min2 and monotonically increasing in the interval [min2, max2], and the mapping value D2 is set to the second pseudo pixel Image data is stored as image data in the image memory in the step 2.4. 7. An original image composed of a plurality of basic pixels is defined as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method in which a pixel to be interpolated between pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. And a second procedure for obtaining the pixel level of the second pseudo pixel and a second procedure for obtaining the pixel level of the second pseudo pixel. 1. Find the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels that are diagonally adjacent to each other.
1 step, and the pixel level average of the four basic pixels, or the four
The 1.2th step of calculating one of the averages of the pixel levels of the two pixels excluding the maximum value max1 and the minimum value min1 from the pixel level of one basic pixel as the average value ave1, and the maximum value max1 and the minimum value min1. The mapping values corresponding to are max1 and
Using a non-linear function that is min1 and monotonically increasing in the interval [min1, max1], the 1.3rd step of obtaining the mapping value D1 of the non-linear function with respect to the average value ave1, and the mapping value D1 is the first pseudo pixel. And a second step of storing the two basic pixels and two first pseudo-pixels, which are the average of two adjacent basic pixels, or the adjacent two basic pixels and two first pseudo pixels. Image data interpolation, characterized in that it comprises a first step 2.1 for calculating one of the averages of the four pixels and a second step 2.2 for storing the average in the image memory as image data of the second pseudo pixel. Method. 8. The first and third steps are to perform a first function F1 which is a non-linear function that monotonically increases in the interval [Min1, Max1] while the minimum and maximum values of the domain and range are both Min1 and Max1. According to the following equation used, the value D1 calculated with the average value ave1 as the input value is the mapping value, and The second step is the following equation using the second function F2, which is a non-linear function that monotonically increases in the interval [Min2, Max2], while the minimum and maximum values of the domain and range are both Min2 and Max2. , The value D2 calculated with the average value ave2 as the input value is the mapping value. The image data interpolation method according to claim 6 or 7, characterized in that. 9. The image data interpolation method according to claim 8, wherein the first function F1 and / or the second function F2 is selected from F (x) in the following equation. [Equation 3] [Equation 4] [Equation 5] 10. An original image composed of a plurality of basic pixels is defined as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolation method in which a pixel to be interpolated between pixels is defined as a second pseudo pixel, and a first pseudo pixel and a second pseudo pixel are arranged in the first image to form a second image. Then, the pixel level of the first pseudo pixel is determined by obtaining the maximum value max1 and the minimum value min1 of the pixel levels of the four basic pixels diagonally adjacent to the first pseudo pixel.
1 step, and the pixel level average of the four basic pixels, or the four
1.2th step of calculating one of the averages of the pixel levels of the two pixels excluding the maximum value max1 and the minimum value min1 from the pixel level of one basic pixel as the average value ave1, the maximum value max1 and the minimum value Calculate the difference with the value min1 and judge that there is a correlation when the difference is smaller than a predetermined value.
Step 1.3, and if there is no correlation, use the average value ave1 as the pixel level of the first pseudo pixel, and if there is correlation, obtain the pixel level of the first pseudo pixel based on a nonlinear function Step 1.4 And a data interpolation method comprising: 11. The pixel level of the second pseudo pixel is determined by determining the maximum pixel level of four pixels, two basic pixels and two first pseudo pixels, which are adjacent to the second pseudo pixel. value
The second step of obtaining max2 and the minimum value min2, the average of the pixel levels of the four pixels, or the maximum value max2 and the minimum value min2 from the pixel levels of the four pixels
The second step of calculating one of the averages of the pixel levels of the two pixels excluding the above as the average value ave2, and the difference between the maximum value max2 and the minimum value min2, and the difference is less than the predetermined value. When it is small, it is judged that there is a correlation.
Step 2.3, and if there is no correlation, use the average value ave2 as the pixel level of the second pseudo pixel, and if there is correlation, obtain the pixel level of the second pseudo pixel based on a non-linear function. Step 2.4 11. The data interpolation method according to claim 10, characterized in that 12. The non-linear function has a domain of input values and a domain of mapping values that match, the input value and its mapping value match at one point within the range, and the first derivative is convex upward. Is. The image data interpolation method according to claim 10 or 11, characterized in that. 13. The image data interpolation method according to claim 12, wherein when the input value is the center value of the domain, the mapping value of the non-linear function is a value near the center of the range. 14. The first image is luminance information which forms a color image together with two color difference information, and the color difference information is interpolated by an average value of a plurality of adjacent pixels. Alternatively, the image data interpolation method described in 9 or 13. 15. The image data interpolation method according to claim 5, 9 or 13, further comprising a third step of reducing an image enlarged by interpolation, wherein the third step is performed in the first direction. Image data characterized by comprising a 3.1 step of reducing the length of each pixel row and a 3.2 step of reducing the length of each pixel row in a second direction orthogonal to the first direction Interpolation method. 16. In the third procedure, the pixel address in the first direction of the image enlarged by interpolation is m, the pixel address in the second direction is n, and the reduction ratio is k (1/2 <k <
Then, the 1st step is to calculate, for each pixel to be obtained in the first direction in the reduced image, the position Pm indicating the luminance level of the corresponding pixel in the reduced image in the image before reduction by the following equation. Pm = (1 / k) m The pixel address m that actually exists on both sides of the position Pm in the first direction.
M0 = int (Pm), which is a sub-step for obtaining 0 and m1 by the following equation, where int is an integer. m1 = m0 + 1 Based on the pixel levels D0 and D1 of the pixels at the addresses m0 and m1, the sub-step to obtain the internal division value by the following formula as the pixel level D and D = (D1-D0) (Pm-m0 ) + D0, and the second step is to calculate the position Pn indicating the brightness level of the pixel after reduction in the image before reduction for each pixel to be obtained in the second direction in the image after reduction using the following formula. Step and Pn = (1 / k) n Address n of the actual pixel on both sides of the position Pn in the second direction
N0 = int (Pn) where 0 and n1 are calculated by the following equation, where int is an integer. n1 = n0 + 1 Based on the pixel levels E0 and E1 of the pixels at addresses n0 and n1, a sub-step that sets the internal value by the following equation as the pixel level E of the pseudo pixel and E = (E1-E0) (Pn- 16. The image data interpolation method according to claim 15, wherein n0) + E0. 17. An original image composed of a plurality of basic pixels is defined as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolating device that forms a second image by arranging a first pseudo pixel and a second pseudo pixel in the first image, with a pixel to be interpolated between pixels as a second pseudo pixel. An image memory for storing the first image and the second image, a maximum value max1 and a minimum value min1 of the pixel levels of the four basic pixels adjacent to the first pseudo pixel, and an average of the four basic pixels. , Or a first average / maximum / minimum determining means for calculating an average value ave1 by calculating an average of two pixels excluding the maximum value max1 and the minimum value min1; and the average value ave1 of the domain and range. A monotonically increasing nonlinear line with both minimum and maximum values of Min1 and Max1. A mapping value D calculated by the following equation using the first function F1 which is a form function
A first function mapping means for storing 1 in the image memory as the image data of the first pseudo pixel; Two base pixels adjacent to the second pseudo pixel and two first pixels
The maximum value max2 and the minimum value min2 of the pixel level of the four pixels of the pseudo pixels, and the average or the maximum value ma of the four pixels.
Average value ave2 which is the average of 2 pixels excluding x2 and minimum value min2
And a second function F2 which is a monotonically increasing nonlinear function in which the minimum and maximum values of the domain and range are both Min2 and Max2 for the average value ave2. Mapping value D calculated by the following equation used
A second function mapping means for storing 2 in the image memory as the image data of the second pseudo pixel; An image data interpolating device comprising: 18. An original image composed of a plurality of basic pixels is defined as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolating device that forms a second image by arranging a first pseudo pixel and a second pseudo pixel in the first image with a pixel to be interpolated between the pixels as a second pseudo pixel. And an image memory for storing the first image and the second image, a maximum value max1 and a minimum value min1 of the pixel levels of the four basic pixels adjacent to the first pseudo pixel, and further the four basic pixels. First average / maximum / minimum determining means for calculating the average value ave1 by calculating either the average value or the average value of two pixels excluding the maximum value max1 and the minimum value min1; and the domain and range of the average value ave1. The minimum and maximum values of are both Min1 and Max1. The mapping value D calculated by the following equation using the first function F1 which is a linear function
A first function mapping means for storing 1 in the image memory as image data of the first pseudo pixel; Either the average of two basic pixels adjacent to the second pseudo pixel or the average of four adjacent pixels of the two basic pixels and the two first pseudo pixels is calculated as an average value, and An image data interpolating device, comprising: an average determining means for storing the average value in the image memory as image data of the second pseudo pixel. 19. The first function F1 and the second function F2.
Is characterized in that the domain of the input value and the range of the mapping value match, the input value and the mapping value match at one point within the range, and the first derivative is convex upward. Item 17. The image data interpolating device according to item 17 or 18. 20. The first function F1 and the second function F2
20. The image data interpolating device according to claim 19, wherein when the input value is the center value of the domain, the mapping value becomes a value near the center of the range. 21. The first function mapping means associates a read address with an input value of the first function, stores a mapping value corresponding to the input, and a maximum value max1 and a minimum value. 21. The image data interpolating apparatus according to claim 20, further comprising a first pixel level determining unit that obtains a mapping value by referring to a first function table based on the value min1 and the average value ave1. 22. The second function mapping means associates a read address with an input value of the second function and stores a mapping value corresponding to the input, and a maximum value max2 and a minimum value. 22. The image data interpolating apparatus according to claim 21, further comprising a second pixel level determining unit that obtains a mapping value by referring to a second function table based on the value min2 and the average value ave2. 23. An original image composed of a plurality of basic pixels is used as a first image, pixels to be interpolated between the diagonally adjacent basic pixels are first pseudo pixels, and the basic pixels vertically and horizontally adjacent to each other. An image data interpolating device that forms a second image by arranging a first pseudo pixel and a second pseudo pixel in the first image with a pixel to be interpolated between the pixels as a second pseudo pixel. And an image memory for storing the first image and the second image, a maximum value max1 and a minimum value min1 of the pixel levels of the four basic pixels adjacent to the first pseudo pixel, and further the four basic pixels. First average / maximum / minimum determining means for calculating an average value ave1 by calculating either the average value or the average value of two pixels excluding the maximum value max1 and the minimum value min1; the maximum value max1 and the minimum value min1; The difference is calculated and the correlation is If there is no correlation, the average value ave1 is set as the pixel level of the first pseudo pixel, and if there is correlation, the first correlation based on the monotonically increasing nonlinear function is used. An image data interpolating device, comprising: first pixel level determining means for determining a pixel level of a pseudo pixel. 24. The image data interpolating device further comprises two base pixels adjacent to the second pseudo pixel and two first pixels.
The maximum value max2 and the minimum value min2 of the pixel level of the four pixels of the pseudo pixels, and the average or the maximum value ma of the four pixels.
Average value ave2 which is the average of 2 pixels excluding x2 and minimum value min2
A second average / maximum / minimum determining means for calculating a difference between the maximum value max2 and the minimum value min2, and a second correlation for determining that there is a correlation when the difference is smaller than a predetermined value. The sex determination means, if there is no correlation, the average value ave2 is set as the pixel level of the second pseudo pixel, and if there is correlation, the pixel level of the second pseudo pixel is calculated based on a monotonically increasing nonlinear function. 24. The image data interpolating device according to claim 23, further comprising two pixel level determining means. 25. The first function F1 and the second function F2
Is characterized in that the domain of the input value and the range of the mapping value match, the input value and the mapping value match at one point within the range, and the first derivative is convex upward. Item 23. The image data interpolating device according to Item 24. 26. The first function F1 and the second function F2
26. The image data interpolating apparatus according to claim 25, wherein when the input value is the center value of the domain, the mapping value becomes a value near the center of the range. 27. The first image is luminance information forming a color image together with two color difference information, and the color difference information is interpolated by an average value of a plurality of adjacent pixels. Or the image data interpolating device according to item 26. 28. The image data interpolating apparatus further reduces the length of each pixel column in the first direction by reducing the length of each pixel row in the image enlarged by the interpolation by the first reducing means. 23. Second reducing means for reducing the length of each pixel column in the second direction orthogonal to the first direction for the captured image.
Or the image data interpolating device according to item 26. 29. When the pixel address in the first direction of the image enlarged by interpolation is m, the pixel address in the second direction is n, and the reduction ratio is k (1/2 <k <1), The first reducing means is, for each pixel to be obtained in the first direction in the reduced image, a position Pm indicating the brightness level of the corresponding pixel of the reduced image Position calculator and Pm = (1 / k) m Address m of the actual pixel on both sides of the position Pm in the first direction
M0 = int (Pm), which is a first adjacent pixel calculation unit that obtains 0 and m1 by the following equation, where int indicates integerization. m1 = m0 + 1 Based on the pixel levels D0 and D1 of the pixels at addresses m0 and m1, the pixel level calculation unit that determines the internally divided value by the following formula as the pixel level D and D = (D1-D0) (Pm -m0) + D0, and the second reduction means calculates, for each pixel to be obtained in the second direction in the reduced image, the position Pn indicating the luminance level of the reduced pixel in the image before reduction as The second position determiner obtained by Pn = (1 / k) n and the address n of the pixel that actually exists on both sides of the position Pn in the second direction.
2nd adjacent pixel calculation part which calculates | requires 0 and n1 by the following formula, and n0 = int (Pn) However, int shows integerization. n1 = n0 + 1 A second pixel level calculation unit that sets the internal division value by the following expression as the pixel level E of the pseudo pixel based on the pixel levels E0 and E1 of the pixels at the addresses n0 and n1, and E = (E1 29. The image data interpolating device according to claim 28, wherein the image data interpolating device comprises -E0) (Pn-n0) + E0.