JP2003230109A - Method of interpolating scanning line, and scanning line converter, and image display - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable accurate oblique correlation type scanning line interpolation by preventing the erroneous detection of oblique correlation, thereby suppressing the deterioration of image quality. <P>SOLUTION: Using the picture elements A-N and a-n on a total of four actual scanning lines consisting of an actual scanning line (n) on a picture element X to be interpolated, an actual scanning line (n+1) under the above picture element X to be interpolated, an actual scanning line (n-1) on the above actual scanning line (n), and an actual scanning line (n+2) under the above actual scanning line (n+1); the direction where the correlation is the highest is detected from among a vertical direction and a plurality of oblique directions centering upon the picture element X to be interpolated, regarding the above picture element X to be interpolated, and the average value of the two picture elements in the above direction where the correlation is the highest on the above actual scanning line (n) and the actual scanning line (n+1) is made the picture element value of the above picture element X to be interpolated. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a scanning line interpolation method for generating an interpolated scanning line between real scanning lines input in a video signal such as a television signal, and more particularly to a correlation of images from a vertical direction and an oblique direction. The present invention relates to a scanning line interpolation method for detecting a direction and generating an interpolation pixel of an interpolation scanning line according to the correlation detection direction. Further, the present invention relates to a scanning line conversion device using such a scanning line interpolation technique, and an image display device equipped with the scanning line conversion device.

[0002]

2. Description of the Related Art A scanning line interpolation technique is used, for example, in a scanning line conversion apparatus for converting an interlaced scanning television signal represented by an NTSC signal into a non-interlaced signal having a double scanning line number. Some image display devices include such a scanning line conversion device and display a signal converted into a non-interlaced signal.

As a scanning line interpolation method, a motion of an input image is detected, and for a still image portion, an interpolated scanning line is generated from an image one field before using a field memory.
For the moving image portion, a motion adaptive type is generally used in which an interpolated scan line is generated from an actual scan line in an input field using a line memory.

FIG. 12 is a block diagram of such a scanning line conversion device. In FIG. 12, 1 is an input terminal of a video signal, 2 and 3 are field memories, 4 is a motion detection circuit, 5 is a line memory, 6 is a double speed conversion circuit, and 7 is an interpolation for a moving image portion which constitutes the double speed conversion circuit 6. A scanning line generation filter, 8 is an interpolating scanning line generation filter for the still image portion which constitutes the double speed conversion circuit 6, 9 is a selector circuit which constitutes the double speed conversion circuit 6, and 10 is a video signal whose scanning line number is doubled. Output terminal.

In the motion detection circuit 4, the difference between the video signal input from the input terminal 1 and the video signal delayed by 2 fields output from the field memory 3 is obtained, and the motion ( Part of the video) is detected.

For the portion detected as a moving image, in the moving image interpolation scanning line generation filter 7, a video signal input from the input terminal 1 and 1 output from the line memory 5 are output.
The video signal delayed by the line, that is, the upper and lower real scan lines in the same field is used to generate the interpolated scan lines between these real scan lines.

In the portion detected as a still image, the video signal delayed by one field output from the field memory 2, that is, the video signal of the previous field is output as it is in the still image portion interpolation scanning line generation filter 8. It is output as an interpolation scan line.

From the interpolation scanning line generation filter 7 for the moving image portion and the interpolation scanning line generation filter 8 for the still image portion, the signal of the actual scanning line and the signal of the interpolation scanning line are output at double speed. According to the motion detection result of the motion detection circuit 4, the selector circuit 9 selects the signal from the moving image interpolation scanning line generation filter 7 in the moving image portion and outputs it to the output terminal 10.
In the still image portion, the signal from the interpolation scanning line generation filter 8 for the still image portion is selected and output to the output terminal 10.

As described above, in the scanning line interpolation, interpolation in the field such as scanning line interpolation in the moving image portion interpolation scanning line generation filter 7 and scanning in the still image portion interpolation scanning line generation filter 8 are performed. There is interpolation between fields such as line interpolation, but the conventional technique described below is scan line interpolation within fields.

FIG. 6 is a diagram for explaining a conventional vertical direction interpolation type scanning line interpolation method using pixels directly above and below the upper and lower actual scanning lines. In FIG. 6, scanning line (n) and scanning line (n + 1) are actual scanning lines, and interpolation scanning line (n) is a scanning line generated from the actual scanning lines. Further, a, b, c are adjacent pixels on the actual scanning line (n), d, e, f are adjacent pixels on the actual scanning line (n + 1), and X is interpolation on the interpolation scanning line (n). It should be a pixel. The pixel b is a pixel directly above the pixel X to be interpolated, and the pixel e is a pixel directly below the pixel X to be interpolated.

In the vertical interpolation type scanning line interpolation method of FIG. 6, the pixel value of the pixel X to be interpolated on the interpolation scanning line (n) is set to the pixel on the scanning line (n) immediately above the pixel X. It is calculated by the average value of b and the pixel e on the scanning line (n + 1) immediately below the pixel X.

However, in the simple method of obtaining the pixel value of the pixel X to be interpolated only from the pixels directly above and below, there is a problem that the oblique line becomes jagged and the image quality deteriorates.

As a scanning line interpolation method for solving such a problem that an oblique line becomes jagged, an oblique scan line in which the correlation between the vertical direction and the oblique direction of the image is detected and the interpolation scan line is generated in the direction having the highest correlation. There is a correlation adaptive scanning line interpolation method.

7 and 8 are diagrams showing a conventional diagonal correlation adaptive scanning line interpolation method. The correlation adaptive scanning line interpolation method described with reference to FIGS.
No. 001-94951. 7 and 8, scanning line (n) and scanning line (n + 1) are actual scanning lines, and interpolation scanning line (n) is a scanning line generated from the actual scanning lines. Further, a, b, c, d, and e are adjacent pixels on the actual scanning line (n), f, g, h, i, and j are adjacent pixels on the scanning line (n + 1), and X is an interpolation scanning line. (N) is a pixel to be interpolated. The pixel c is a pixel directly above the pixel X to be interpolated, and the pixel h is a pixel directly below the pixel X to be interpolated.

In the conventional diagonal correlation adaptive scanning line interpolation method shown in FIGS. 7 and 8, first, as shown in FIG.
Absolute value | c−h | of the difference between the pixel a and the pixel h
Absolute value | a−i | of the difference between pixel b and pixel j, absolute value | b−j | of the difference between pixel b and pixel j, absolute value | d−f | of the difference between pixel d and pixel f, and pixel e and pixel absolute value of the difference from g | e−g
|, Absolute value | b−h | of the difference between pixel b and pixel h, absolute value | c−i | of the difference between pixel c and pixel i, absolute value | c− of the difference between pixel c and pixel g g | and the absolute value | d−h | of the difference between the pixel d and the pixel h are calculated.

Then, the correlation in the vertical direction is {2 × | c−h
|}, The first diagonal correlation is {| a-i | + | b-j
│}, the second diagonal correlation is {| d−f | + | e−g
|, The third diagonal correlation is {| b-h | + | c-i
|}, The correlation in the fourth diagonal direction is {| c-g | + | d-h
|, And the direction with the smallest value is the direction with the highest correlation.

When the direction having the highest correlation is the vertical direction, the average value of the pixels c and h is used as the pixel value of the pixel X to be interpolated, as shown in FIG.

When the direction having the highest correlation is the first diagonal direction, the average value of the pixel a and the pixel j is set as the pixel value of the pixel X to be interpolated, as shown in FIG.

When the direction having the highest correlation is the second diagonal direction, the average value of the pixel e and the pixel f is set as the pixel value of the pixel X to be interpolated, as shown in FIG.

When the direction having the highest correlation is the third diagonal direction, the average value of the pixel b and the pixel i is set as the pixel value of the pixel X to be interpolated, as shown in FIG.

When the direction having the highest correlation is the fourth diagonal direction, the average value of the pixel d and the pixel g is set as the pixel value of the pixel X to be interpolated, as shown in FIG.

As described above, in the conventional diagonal correlation adaptive scanning line interpolation method, the direction having the highest correlation is detected from the vertical direction and a plurality of diagonal directions, and the pixel value of the pixel X to be interpolated is obtained according to the direction. As a result, the diagonal line can be smoothed.

[0023]

However, in the above-mentioned conventional diagonal correlation adaptive scanning line interpolation method, there is a possibility that the direction of the highest correlation may be erroneously detected in the input image representing a part of the cross. There is a problem in that the image quality is deteriorated when is incorrect.

9 to 11 are views for explaining the problems of the conventional diagonal correlation adaptive scanning line interpolation method. FIG. 9 is a diagram of an input image representing a part of a cross, and FIG. 9 is an interpolated image of the image of FIG. 9 generated by the oblique correlation adaptive scanning line interpolation method of FIG. 11, and FIG. 11 is an ideal interpolated image of the image of FIG. 9 to 11, the scanning line (n-1), the scanning line (n), the scanning line (n + 1), and the scanning line (n + 2) are real scanning lines, and the interpolation scanning line (n-
1), the interpolation scan line (n), and the interpolation scan line (n + 1) are scan lines generated from the actual scan line. Also, a, b, c,
d and e are adjacent pixels on the actual scanning line (n), f, g,
h, i, and j are adjacent pixels on the scanning line (n + 1), and X is a pixel to be interpolated. White circles, black circles, and gray squares are pixels on the actual scanning line or the interpolation scanning line. White circles and black circles are pixels having different pixel values, and white circles and black circles have substantially the same pixel value. The gray square pixel value is the average value of the white circle pixel values and the black circle pixel values.

The image showing a part of the cross in FIG. 9 corresponds to, for example, a portion where the vertical line and the horizontal line of the crosshatch signal intersect. When the correlation of the position of the pixel X to be interpolated is detected in the input image of FIG. 9 by the conventional diagonal correlation adaptive scanning line interpolation method, as shown by the arrow in FIG. It is detected that the correlation in the diagonal direction connecting the pixel f) is the highest, and the pixel value of the pixel X becomes the same as that of the black circle pixel.

Then, when the input image of FIG. 9 is interpolated by the above-mentioned conventional diagonal correlation adaptive scanning line interpolation method, the result is as shown in FIG.
Gray square pixels, which are the average pixel values of white circles and black circles, are interpolated, and the root portion of the horizontal line of the cross is interpolated by black circle pixels. As described above, when the input image representing a part of the cross in FIG. 9 is interpolated by the conventional diagonal correlation adaptive scanning line interpolation method, an unnatural image as shown in FIG. 10 is obtained.

For an input image representing a part of a cross as shown in FIG. 9, the direction having the highest correlation is set as the vertical direction, and the pixel value of the interpolation pixel is set to the true value as in the conventional vertical interpolation type scanning line interpolation method. Ideally, the average value of the upper and lower pixels is used, and as shown in FIG. 11, it is ideal to interpolate from the tip of the horizontal line of the cross to the root with gray square pixels obtained by averaging the pixel values of the white and black circles. .

The present invention has been made to solve such a conventional problem, and it is possible to suppress the image quality deterioration due to the false detection of the diagonal correlation and enable the highly accurate diagonal correlation type scanning line interpolation. To aim.

[0029]

A scanning line interpolation method according to claim 1 of the present invention is a diagonal interpolation adaptive scanning line interpolation method, wherein at least a first actual scanning line above a pixel to be interpolated is used. , Using a pixel on a total of two actual scanning lines of the second actual scanning line below the pixel to be interpolated, from the vertical direction and a plurality of diagonal directions centering on the pixel to be interpolated,
A correlation detecting step of detecting the direction having the highest correlation among the pixels to be interpolated, by giving priority to the vertical direction first and sequentially giving priority to diagonal directions as close as possible to the vertical direction;
And 2 in the direction with the highest correlation on the second actual scan line.
And a step of setting an average value of pixels as a pixel value of the pixel to be interpolated.

A scanning line interpolation method according to a second aspect is the scanning line interpolation method according to the first aspect, wherein in the correlation detecting step, at least the first and second actual scanning lines are on the respective actual scanning lines. It is characterized by using seven pixels to detect the direction with the highest correlation from a total of seven directions, namely, the vertical direction, three types of diagonal directions to the right and three types of diagonal directions to the left. is there.

According to a third aspect of the present invention, there is provided an oblique correlation adaptive scanning line interpolation method in which at least a first actual scanning line above a pixel to be interpolated and a pixel below the pixel to be interpolated. On a total of four real scanning lines, a second real scanning line, a third real scanning line above the first real scanning line, and a fourth real scanning line below the second real scanning line. Correlation detection step of detecting the direction having the highest correlation with respect to the pixel to be interpolated from among the vertical direction and a plurality of diagonal directions centering on the pixel to be interpolated, using The step of setting the average value of the two pixels in the direction having the highest correlation on the two actual scanning lines as the pixel value of the pixel to be interpolated.

A scanning line interpolation method according to a fourth aspect is the scanning line interpolation method according to the third aspect, wherein in the correlation detecting step, the interpolation is performed using at least pixels on the first and second actual scanning lines. The direction having the highest correlation with respect to the pixel to be interpolated is detected from the vertical direction and a plurality of diagonal directions centered on the pixel to be interpolated, and at least the pixels on the first and third actual scanning lines are used, The direction having the highest correlation with respect to the first reference pixel from the vertical direction and the plurality of diagonal directions centering on the first reference pixel between the first actual scanning line and the third actual scanning line. In the vertical direction centered on the second reference pixel between the second real scanning line and the fourth real scanning line, using at least the pixels on the second and fourth real scanning lines. And from multiple diagonal directions, The direction having the highest correlation with respect to the second reference pixel is detected, and the direction to be interpolated and the directions detected with the highest correlation with respect to the first and second reference pixels are referred to. The feature is that the direction having the highest correlation is determined for the pixel to be interpolated.

A scanning line interpolation method according to a fifth aspect is the scanning line interpolation method according to the third or fourth aspect, wherein in the correlation detecting step, at least the actual scanning lines from the first to the fourth are respectively detected. Using 7 pixels on the actual scanning line, the direction with the highest correlation is detected from a total of 7 types of directions including vertical direction, 3 types of diagonal directions to the right and 3 types of diagonal directions to the left. It is what

According to a sixth aspect of the present invention, in the scanning line converting apparatus, at least a first actual scanning line above the pixel to be interpolated and a second actual scanning line below the pixel to be interpolated. Using a total of two pixels on the actual scanning line, from the vertical direction and a plurality of diagonal directions centering on the pixel to be interpolated,
On the first and second actual scanning lines, the correlation detection means detects the direction having the highest correlation among the pixels to be interpolated by giving priority to the vertical direction first and sequentially giving priority to diagonal directions as close as possible to the vertical direction. It is characterized by comprising an interpolation scanning line generating means for generating an interpolation scanning line by using an average value of the two pixels in the direction having the highest correlation as a pixel value of the pixel to be interpolated.

According to a seventh aspect of the present invention, there is provided a scanning line converting apparatus which has at least a first actual scanning line above a pixel to be interpolated, a second actual scanning line below the pixel to be interpolated, and the first actual scanning line. Pixels on a total of four real scanning lines including the third real scanning line above the real scanning line and the fourth real scanning line below the second real scanning line Correlation detecting means for detecting a direction having the highest correlation with respect to the pixel to be interpolated, from among the vertical direction and the plurality of diagonal directions centered on
And 2 in the direction with the highest correlation on the second actual scan line.
Interpolation scanning line generating means for generating an interpolation scanning line using the average value of the pixels as the pixel value of the pixel to be interpolated is provided.

The scanning line converting apparatus according to claim 8 is the scanning line converting apparatus according to claim 6 or 7, wherein the interpolation scanning line generating means is a moving image interpolation scanning line generating means, and Motion detection means for detecting the motion of the video signal, still image interpolation scanning line generation means, and a signal output from the moving image interpolation scanning line generation means based on the result detected by the motion detection means and the above It is characterized by further comprising: selecting means for selecting one of the signals output from the still image interpolation scanning line generating means.

An image display device according to a ninth aspect of the present invention is characterized by including the scanning line conversion device according to any one of the sixth to eighth aspects.

[0038]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1. In the following first embodiment, 5 pixels on the actual scanning line above the pixel to be interpolated on the interpolation scanning line and 5 pixels on the actual scanning line below the pixel to be interpolated.
Using a total of 10 pixels, the direction with the highest correlation is detected from a total of 5 different directions of the vertical direction and 4 types of diagonal directions, and the pixel of the pixel to be interpolated according to the detected correlation direction. A method for obtaining the value will be described. In the following description, the diagonal correlation adaptive scanning line interpolation method according to the first embodiment will be described.
It shall be applied to the interpolation scanning line generation filter 7 for moving image parts of FIG.

FIG. 2 is a diagram for explaining the diagonal correlation adaptive scanning line interpolation method according to the first embodiment. In FIG. 2, scanning line (n) and scanning line (n + 1) are actual scanning lines, and interpolation scanning line (n) is a scanning line generated from the actual scanning lines. B, C, D, E, and F are adjacent pixels on the actual scanning line (n), b, c, d, e, and f are adjacent pixels on the actual scanning line (n + 1), and X is interpolation scanning. Pixels to be interpolated on line (n). The pixel D is a pixel directly above the pixel X to be interpolated, and the pixel d is a pixel directly below the pixel X to be interpolated.

In the first embodiment, the pixel X to be interpolated
In order to obtain the pixel value of, the direction with the highest correlation is detected from a total of 5 types of directions, the vertical direction and the four types of diagonal directions. The detection procedure will be described below.

In the following description of the first embodiment, the direction connecting the pixel B and the pixel f is the oblique direction Bf, the direction connecting the pixel C and the pixel e is the oblique direction Ce, and the pixel E and the pixel c are the same. The connecting direction is an oblique direction Ec, and the connecting direction of the pixel F and the pixel b is an oblique direction Fb. The direction connecting the pixel D and the pixel d is the vertical direction.

First, as shown in FIG. 2, pixel B and pixel f
Absolute value | B−f |, absolute value | C−e | of pixel C and pixel e, and absolute value | D− of pixel D and pixel d.
d |, the absolute value | E−c | of the difference between the pixel E and the pixel c, and the absolute value | F−b | of the difference between the pixel F and the pixel b are calculated. That is, the absolute values of the differences between the pixels in the vertical direction and the four types of diagonal directions with the pixel X to be interpolated as the center are calculated.

Next, a sum SL of absolute values of the difference in the upward-sloping diagonal direction SL = {| Bf | + | C-e |} and a sum total of absolute values of the difference in the upward-sloping diagonal direction SR = {| E −c | + | F−b |}, and the absolute value of the difference between these sums | SL-S
R | is obtained, and this | SL-SR | is compared with a fixed value α.

If | SL-SR | <α, it is determined that the correlation in the diagonal direction is low, and the vertical direction is the direction with the highest correlation.

If | SL-SR | ≧ α, the sum S
L and SR are compared, and if SL <SR, the one having the smallest absolute value of the difference between the two upward diagonal directions Bf and Ce is set as the highly correlated candidate Δ. The one having the smallest absolute value of the difference between the upwardly inclined directions Ec and Fb is set as the candidate Δ having a high correlation. When there are equal values, the diagonal direction close to the vertical direction is prioritized. Then, the above-mentioned candidate Δ is compared with the absolute value of the vertical difference | D−d |, and if Δ <| D−d |, the diagonal direction of the above-mentioned candidate Δ is directly set as the direction with the highest correlation, If Δ ≧ | D−d |, the vertical direction is the direction with the highest correlation.

That is, | SL-SR | ≧ α and SL <S
If R, the absolute value of the difference in the diagonal direction Bf | B−f |
The smaller of the two absolute values | C−e | of the difference in the oblique direction Ce is selected as the candidate Δ having the highest correlation. When both absolute values are equal, the oblique direction Ce is prioritized. Then, the above-mentioned candidate Δ1 is compared with the absolute value of the vertical difference | D−d |, and if Δ <| D−d |, the diagonally upward-sloping direction selected as Δ is the direction with the highest correlation. If Δ ≧ | D−d |, the vertical direction is the direction with the highest correlation.

Similarly, | SL-SR | ≧ α and SL
If ≧ SR, the absolute value of the difference in the diagonal direction Ec | E−c |
And the smaller absolute value | F−b | of the difference in the diagonal direction Fb, the smaller one is selected as the candidate Δ having the highest correlation.
When both absolute values are equal, the diagonal direction Ec is prioritized. Then, the above-mentioned candidate Δ is compared with the absolute value of the vertical difference | D−d |, and if Δ <| D−d |, the upward-sloping diagonal direction selected as Δ is the direction with the highest correlation. If Δ ≧ | D−d |, the vertical direction is the direction with the highest correlation.

In the first embodiment, the pixel value of the pixel X to be interpolated is calculated in accordance with the direction detected to have the highest correlation from the total of 5 types of directions including the vertical direction and the four types of diagonal directions. The calculation procedure will be described below.

When the vertical direction is detected as the direction having the highest correlation, the average value {(D
Let + d) / 2} be the pixel value of the pixel X to be interpolated.

When the diagonal direction Bf is detected as the direction having the highest correlation, the average value {(B + f) / 2} of the pixel B and the pixel f in FIG. 2 is set as the pixel value of the pixel X to be interpolated. .

When the oblique direction Ce is detected as the direction having the highest correlation, the average value {(C + e) / 2} of the pixel C and the pixel e in FIG. 2 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Ec is detected as the direction having the highest correlation, the average value {(E + c) / 2} of the pixel E and the pixel c in FIG. 2 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Fb is detected as the direction having the highest correlation, the average value {(F + b) / 2} of the pixel F and the pixel b in FIG. 2 is set as the pixel value of the pixel X to be interpolated. .

As described above, according to the first embodiment, the vertical direction is given the highest priority out of the total of 5 directions including the vertical direction and the four oblique directions, and the oblique directions that are as close to the vertical direction as possible are sequentially prioritized. Then, by detecting the direction with the highest correlation, it is possible to reduce false detection of diagonal correlation and perform highly accurate interpolation.

Embodiment 2. In the following second embodiment,
Similar to the first embodiment, a total of 14 pixels are used, that is, 7 pixels on the actual scanning line above the pixel to be interpolated on the interpolation scanning line and 7 pixels on the actual scanning line below the pixel to be interpolated. Next, a method of detecting the direction having the highest correlation from the total of seven types of directions, namely the vertical direction and the six types of diagonal directions, and obtaining the pixel value of the pixel to be interpolated according to the detected correlation direction will be described. In the following description, the diagonal correlation adaptive scanning line interpolation method of the first embodiment is applied to the moving image portion interpolation scanning line generation filter 7 of FIG.

FIG. 3 is a diagram for explaining the diagonal correlation adaptive scanning line interpolation method according to the second embodiment. In FIG. 3, scanning line (n) and scanning line (n + 1) are actual scanning lines, and interpolation scanning line (n) is a scanning line generated from the actual scanning lines. Further, A, B, C, D, E, F, and G are adjacent pixels on the actual scanning line (n), and a, b, c, d, e, f, and g are on the actual scanning line (n + 1). Adjacent pixels, X, are pixels to be interpolated on the interpolation scanning line (n). The pixel D is a pixel directly above the pixel X to be interpolated, and the pixel d is a pixel directly below the pixel X to be interpolated.

In the second embodiment, the pixel X to be interpolated
In order to obtain the pixel value of, the direction with the highest correlation is detected from the total of 7 types of directions including the vertical direction and 6 types of diagonal directions. The detection procedure will be described below.

In the following description of the second embodiment, the direction connecting the pixel A and the pixel g is the oblique direction Ag, the direction connecting the pixel B and the pixel f is the oblique direction Bf, and the pixel C and the pixel e are the same. The connecting direction is oblique direction Ce, the direction connecting pixel E and pixel c is oblique direction Ec, the direction connecting pixel F and pixel b is oblique direction Fb, and the direction connecting pixel G and pixel a is oblique. The direction is Ga. The direction connecting the pixel D and the pixel d is the vertical direction.

First, as shown in FIG. 3, pixel A and pixel g
Absolute value | A−g |, absolute value of difference between pixel B and pixel f | B−f |, absolute value of difference between pixel C and pixel e | C−
e |, absolute value | D−d | of the difference between pixel D and pixel d, absolute value | E−c | of the difference between pixel E and pixel c, pixel F and pixel b
The absolute value | F−b | of the difference and the absolute value | G−a | of the difference between the pixel G and the pixel a are calculated. That is, the absolute values of the differences between the pixels in the vertical direction and the six types of diagonal directions with the pixel X to be interpolated as the center are calculated.

Next, the sum SL of absolute values of the difference in the upward-sloping diagonal direction SL = {| A-g | + | B-f | + | C-e |} and the absolute value of the difference in the upward-sloping diagonal direction Sum SR = {| E-c | + | F-b | + | G-a |} is obtained, and the absolute value of the difference between these sums | SL-S
R | is obtained, and this | SL-SR | is compared with a fixed value α.

If | SL-SR | <α, it is determined that the correlation in the diagonal direction is low, and the vertical direction is the direction with the highest correlation.

If | SL-SR | ≧ α, the sum S
L and SR are compared, and if SL <SR, the one having the smallest absolute value of the difference among the three upward-sloping diagonal directions Ag, Bf, Ce is set as the highly correlated candidate Δ, and if SL ≧ SR,
The one having the smallest absolute value of the difference among the three upward-sloping diagonal directions Ec, Fb, and Ga is set as the candidate Δ having a high correlation. When there are equal values, the diagonal direction close to the vertical direction is prioritized. Then, the above-mentioned candidate Δ is compared with the absolute value of the vertical difference | D−d |, and if Δ <| D−d |, the diagonal direction of the above-mentioned candidate Δ is directly set as the direction with the highest correlation, If Δ ≧ | D−d |, the vertical direction is the direction with the highest correlation.

That is, | SL-SR | ≧ α and SL <S
If R, the absolute value of the difference in the diagonal direction Ag | A−g |
Absolute value | B−f | of the difference in diagonal direction Bf and diagonal direction C
Among the three absolute values | C−e | of the difference of e, the smallest one is selected as the candidate Δ having the highest correlation. When there are equal values, the diagonal direction Ce, the diagonal direction Bg, and the diagonal direction Ag are given priority in this order. Then, the above candidate Δ is compared with the absolute value of the vertical difference | D−d |, and Δ <| D
If -d |, the diagonal direction selected as Δ is directly regarded as the candidate with the highest correlation, and if Δ ≧ | D−d |, the vertical direction is determined as the direction with the highest correlation.

Similarly, | SL-SR | ≧ α and SL ≧ S
If R, the absolute value of the difference in the diagonal direction Ec | E-c |
Absolute value | F−b | of the difference in diagonal direction Fb and diagonal direction G
The smallest of the absolute values | G−a | of the difference of a is selected as the candidate Δ having the highest correlation. If there are equal values, the diagonal direction Ec, the diagonal direction Fb, and the diagonal direction Ga are given priority in this order. Then, the above candidate Δ is compared with the absolute value of the vertical difference | D−d |, and Δ <| D
If -d |, the diagonal direction selected as Δ is directly regarded as the candidate with the highest correlation, and if Δ ≧ | D−d |, the vertical direction is determined as the direction with the highest correlation.

In the second embodiment, the pixel value of the pixel X to be interpolated is calculated in accordance with the direction detected to have the highest correlation from the total of 7 types of directions including the vertical direction and the 6 types of diagonal directions. The calculation procedure will be described below.

When the vertical direction is selected as the direction having the highest correlation, the average value {(D
Let + d) / 2} be the pixel value of the pixel X to be interpolated.

When the diagonal direction Ag is detected as the direction having the highest correlation, the average value {(A + g) / 2} of the pixel A and the pixel g in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Bf is detected as the direction having the highest correlation, the average value {(B + f) / 2} of the pixel B and the pixel f in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

When the oblique direction Ce is detected as the direction having the highest correlation, the average value {(C + e) / 2} of the pixel C and the pixel e in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Ec is detected as the direction having the highest correlation, the average value {(E + c) / 2} of the pixel E and the pixel c in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Fb is detected as the direction having the highest correlation, the average value {(F + b) / 2} of the pixel F and the pixel b in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

When the diagonal direction Ga is detected as the direction having the highest correlation, the average value {(G + a) / 2} of the pixel G and the pixel a in FIG. 3 is set as the pixel value of the pixel X to be interpolated. .

As described above, according to the second embodiment, the vertical direction is given the highest priority from the total of 7 kinds of directions including the vertical direction and 6 kinds of oblique directions, and the oblique direction which is next as close to the vertical direction as possible is selected. By sequentially prioritizing and detecting the direction with the highest correlation, it is possible to reduce false detection of diagonal correlation and perform interpolation with higher accuracy than in the first embodiment.

Third Embodiment In the following third embodiment,
By using two real scanning lines above the pixel to be interpolated on the interpolation scanning line and two real scanning lines below the pixel to be interpolated, a total of four pixels on the real scanning line, A method of detecting the direction with the highest correlation from the total of 7 types of 6 types of diagonal directions and obtaining the pixel pixel value to be interpolated according to the detected correlation direction will be described.

As described above, since four real scanning lines are used in the third embodiment, four real scanning lines are used in the interpolation scanning line generation filter for the moving image section of the scanning line conversion apparatus to which the third embodiment is applied. The signal of the line is input, the direction having the highest correlation is detected using the pixels of these four actual scanning lines, and the interpolated pixel is generated.

FIG. 13 is a block diagram of a scanning line conversion apparatus to which the diagonal correlation adaptive scanning line interpolation method of the third embodiment is applied. In FIG. 13, the same or corresponding parts as those in FIG. 12 are designated by the same reference numerals, 1 is a video signal input terminal, 2 and 3 are field memories, 4 is a motion detection circuit, and 5, 11 and 12 are A line memory, 6 is a double speed conversion circuit, 7 is an interpolation scanning line generation filter for a moving image portion which constitutes the double speed conversion circuit 6, 8 is an interpolation scanning line generation filter for a still image portion which constitutes the double speed conversion circuit 6, and 9 is Double speed conversion circuit 6
The selector circuit 10 and 10 are output terminals for a video signal whose number of scanning lines is doubled.

The scanning line converting apparatus of FIG. 13 is different from the scanning line converting apparatus of FIG. 12 in that line memories 11 and 12 are provided.
The output signals of the line memories 11 and 12 are input to the moving image interpolation filter 7. Therefore, in the scanning line conversion apparatus of FIG. 13, the video signal input to the input terminal 1, the output signal from the line memory 5 which is the input signal delayed by one line, and the line memory 11 which is the input signal delayed by two lines. Output signal from and input signal 3
The signals of a total of four actual scanning lines, which are the output signals from the line memory 12 delayed by the line, are input to the moving image interpolation filter 7. In the following description, the diagonal correlation adaptive scanning line interpolation method of the third embodiment is applied to the moving image portion interpolation scanning line generation filter 7 of FIG.

FIG. 1 is a diagram for explaining the diagonal correlation adaptive scanning line interpolation method according to the third embodiment. In FIG. 1, scanning line (n-1), scanning line (n), scanning line (n + 1), scanning line (n + 2) are real scanning lines, and interpolation scanning line (n-1),
The interpolation scanning line (n) and the interpolation scanning line (n + 1) are scanning lines generated from the actual scanning line. Also, A, B, C, D,
E, F, and G are adjacent pixels on the actual scanning line (n), a,
b, c, d, e, f, g are adjacent pixels on the real scanning line (n + 1), and H, I, J, K, L, M, N are real scanning lines (n + 1).
-1) Adjacent pixels on h, i, j, k, l, m, n
Is an adjacent pixel on the actual scanning line (n + 2), X is a pixel to be interpolated on the interpolation scanning line (n), and X1 is an interpolation scanning line (n−).
The reference pixel on 1) and X2 are the reference pixels on the interpolation scanning line (n + 1). The pixel D is a pixel directly above the pixel X to be interpolated, the reference pixel X1 is a pixel directly above the pixel D, and the pixel K is a pixel directly above the reference pixel X1. Also,
The pixel d is a pixel directly below the pixel X to be interpolated, the reference pixel X2 is a pixel directly below the pixel d, and the pixel k is a pixel directly below the reference pixel X2.

In the third embodiment, the pixel X to be interpolated
In order to obtain the pixel value of, the direction with the highest correlation in the pixel X to be interpolated and the reference pixels X1 and X2 is detected from each of the total of seven types of directions including the vertical direction and six types of diagonal directions, and interpolation is performed. The direction with the highest correlation for the pixel X to be interpolated and the reference pixels X1 and X2 is referred to, and the direction with the highest correlation for the pixel X to be interpolated is determined. The detection procedure will be described below.

In the following description of the third embodiment, the direction connecting the pixel A and the pixel g is the oblique direction Ag, the direction connecting the pixel B and the pixel f is the oblique direction Bf, and the pixel C and the pixel e are the same. The connecting direction is oblique direction Ce, the direction connecting pixel E and pixel c is oblique direction Ec, the direction connecting pixel F and pixel b is oblique direction Fb, and the direction connecting pixel G and pixel a is oblique. Let Ga be the direction Ga, the direction connecting the pixels H and G is the diagonal direction HG, and the direction connecting the pixels I and F is the diagonal direction I.
F, the direction connecting the pixels J and E is a diagonal direction JE, the direction connecting the pixels L and C is a diagonal direction LC, the direction connecting the pixels M and B is a diagonal direction MB, and the pixel N. Pixel A
Is a diagonal direction NA, a direction connecting the pixel a and the pixel n is a diagonal direction an, a direction connecting the pixel b and the pixel m is a diagonal direction bm, and a direction connecting the pixel c and the pixel l. The direction is diagonal direction cl, the direction connecting pixel e and pixel j is diagonal direction ej, and the direction connecting pixel f and pixel i is diagonal direction f.
i, the direction connecting the pixel g and the pixel h is defined as an oblique direction gh. Further, the oblique directions Ag, HG, an are the first oblique direction, the oblique directions Bf, IF, bm are the second oblique directions, and the oblique directions Ce, JE, cl are the third oblique direction, and the oblique direction E.
c, LC, ej in the fourth diagonal direction, diagonal direction Fb, M
B, fi in the fifth diagonal direction, diagonal directions Ga, NA, gh
Is the sixth diagonal direction. The direction connecting the pixel D and the pixel d, the direction connecting the pixel K and the pixel D, and the direction connecting the pixel d and the pixel k are vertical directions.

First, similarly to the second embodiment, the pixels on the actual scanning lines above and below the pixel X to be interpolated are used to measure the pixel X to be interpolated in the vertical direction and the first to sixth diagonal directions. The direction with the highest correlation is detected among the seven types of directions.

Similarly, by using the pixels on the actual scanning lines above and below the interpolation scanning line (n-1), the reference pixel X1 can be selected from the vertical direction and the first to sixth diagonal directions in a total of seven directions. The direction with the highest correlation is detected.

Similarly, using the pixels on the actual scanning lines above and below the interpolation scanning line (n + 1), the reference pixel X2 is the most vertical direction and the first to sixth oblique directions among the total of seven kinds of directions. Detect the direction with high correlation.

That is, for the pixel X to be interpolated, the pixels A, B, C, D, E, F, G on the actual scanning line (n) and the pixels a, b, c on the actual scanning line (n + 1). , D, e, f, g
And the absolute value of the difference | A−g |, | B−f |, |
C-e |, | D-d |, | E-c |, | F-b |, | G
−a | is calculated respectively, and the sum SL of absolute values of the difference in the upward-sloping diagonal direction SL = {| A−g | + | B−f | + | C−e |} and the absolute difference in the upward-sloping diagonal direction are calculated. The sum of the values SR = {| E−c | + | F−b | + | G−a |} is obtained, and the absolute value of the difference between these sums | SL−S
R | is obtained, and this | SL-SR | is compared with a fixed value α.

If | SL-SR | <α, it is determined that the correlation in the diagonal direction is low, and the vertical direction is set as the direction having the highest correlation for the pixel X to be interpolated.

Also, | SL-SR | ≧ α and SL <SR
Then, the absolute value | A−g | of the difference in the diagonal direction Ag, the absolute value | B−f | of the difference in the diagonal direction Bf, and the diagonal direction Ce
The smallest of the absolute values | C−e | of the difference of 3 is selected as the candidate Δ having the highest correlation. When there are equal values, the diagonal direction Ce, the diagonal direction Bg, and the diagonal direction Ag are given priority in this order. Then, the above candidate Δ is compared with the absolute value of the vertical difference | D−d |, and Δ <| D−
If d |, the diagonal direction selected for Δ is the direction with the highest correlation for the pixel X to be interpolated, and Δ ≧ | D−d
If |, the vertical direction is the direction with the highest correlation for the pixel X to be interpolated.

Further, | SL-SR | ≧ α and SL ≧ SR
Then, the absolute value | E−c | of the difference in the diagonal direction Ec, the absolute value | F−b | of the difference in the diagonal direction Fb, and the diagonal direction Ga
The smallest one among the three absolute values | G−a | of the difference is selected as the candidate Δ having the highest correlation. If there are equal values, the diagonal direction Ec, the diagonal direction Fb, and the diagonal direction Ga are given priority in this order. Then, the above candidate Δ is compared with the absolute value of the vertical difference | D−d |, and Δ <| D−
If d |, the diagonal direction selected for Δ is the direction with the highest correlation for the pixel X to be interpolated, and Δ ≧ | D−d
If |, the vertical direction is the direction with the highest correlation for the pixel X to be interpolated.

For the reference pixel X1, the actual scanning line (n-
1) Using the pixels H, I, J, K, L, M, N on the real scanning line and the pixels A, B, C, D, E, F, G on the actual scanning line (n), the absolute value of the difference | H-G |, | I-F |, | J-E
|, | K-D |, | L-C |, | M-B |, | NA- |
Is calculated, and the sum of absolute values of the difference in the leftward diagonal direction SL1 = {| HG | + | IF | + | J-E |}, and the sum of the absolute value of the difference in the diagonal direction to the right. SR1 = {| LC | + | MB | + | NA |}, and the absolute value of the difference between these sums | SL1-
SR1 | is obtained, and this | SL1-SR1 | is compared with the fixed value α.

If | SL1-SR1 | <α,
It is determined that the correlation in the diagonal direction is low, and the vertical direction is set as the direction having the highest correlation with respect to the reference pixel X1.

Also, | SL1-SR1 | ≧ α and SL1
<If SR1, absolute value of difference in diagonal direction HG | HG
|, The absolute value | I−F | of the difference in the diagonal direction IF, and the absolute value | J−E | of the difference in the diagonal direction JE are selected as the candidate Δ1 having the highest correlation. In addition,
If they are equal, the diagonal direction close to the vertical direction is given priority. That is, the diagonal direction JE, the diagonal direction IF, and the diagonal direction HG are prioritized in this order. Then, the above candidate Δ1
And the absolute value of the vertical difference | K−D |
<| K−D |, the diagonal direction selected as Δ1 is set as the direction having the highest correlation with respect to the reference pixel X1, and Δ1 ≧
If | K−D |, the vertical direction is the direction having the highest correlation with respect to the reference pixel X1.

Also, | SL1-SR1 | ≧ α and SL1
If ≧ SR1, the absolute value of the difference in the diagonal direction LC | LC
|, The absolute value | MB of the difference in the diagonal direction MB, and the absolute value | NA of the difference in the diagonal direction NA | NA, the smallest one is selected as the candidate Δ1 having the highest correlation. In addition,
If they are equal, the diagonal direction close to the vertical direction is given priority. That is, the oblique direction LC, the oblique direction MB, and the oblique direction NA are prioritized in this order. Then, the above candidate Δ1
And the absolute value of the vertical difference | K−D |
<| K−D |, the diagonal direction selected as Δ1 is set as the direction having the highest correlation with respect to the reference pixel X1, and Δ1 ≧
If | K−D |, the vertical direction is the direction having the highest correlation with respect to the reference pixel X1.

For the reference pixel X2, the actual scanning line (n +
1) Using pixels a, b, c, d, e, f, and g on the pixels and pixels h, i, j, k, l, m, and n on the actual scanning line (n + 2), the absolute value of the difference | A-n |, | b-m |, | c-l
|, | D-k |, | e-j |, | f-i |, | g-h |
Respectively, the sum of absolute values of the difference in the leftward diagonal direction SL2 = {| a−n | + | b−m | + | c−1 |}, and the sum of absolute values of the difference in the upward diagonal direction. SR2 = {| e−j | + | f−i | + | g−h |}, and the absolute value of the difference between these sums | SL2-
SR2 | is obtained, and this | SL2-SR2 | is compared with the fixed value α.

If | SL2-SR2 | <α,
It is determined that the correlation in the diagonal direction is low, and the vertical direction is the direction having the highest correlation with respect to the reference pixel X2.

Also, | SL2-SR2 | ≧ α and SL2
<SR2, absolute value of difference in diagonal direction an | a−n
|, The absolute value | b−m | of the difference in the diagonal direction bm, and the absolute value | c−l | of the difference in the diagonal direction cl, the smallest one is selected as the candidate Δ2 having the highest correlation. In addition,
If they are equal, the diagonal direction close to the vertical direction is given priority. That is, the slanting direction cl, the slanting direction bm, and the slanting direction an are given priority in this order. Then, the above candidate Δ2
And the absolute value of the vertical difference | d−k | are compared, and Δ2
<| D−k |, the diagonal direction selected as Δ2 is set as the direction having the highest correlation with respect to the reference pixel X2, and Δ2 ≧
If | d−k |, the vertical direction is the direction having the highest correlation with respect to the reference pixel X2.

Also, | SL2-SR2 | ≧ α and SL2
If ≧ SR2, the absolute value of the difference in the diagonal direction ej | e−j
|, The absolute value | f−i | of the difference in the diagonal direction fi and the absolute value | g−h | of the difference in the diagonal direction gh are selected as the candidate Δ2 having the highest correlation. In addition,
If they are equal, the diagonal direction close to the vertical direction is given priority. That is, the oblique direction ej, the oblique direction fi, and the oblique direction gh are prioritized in this order. Then, the above candidate Δ2
And the absolute value of the vertical difference | d−k | are compared, and Δ2
<| D−k |, the diagonal direction selected as Δ2 is set as the direction having the highest correlation with respect to the reference pixel X2, and Δ2 ≧
If | d−k |, the vertical direction is the direction having the highest correlation with respect to the reference pixel X2.

In the third embodiment, the pixel value of the pixel X to be interpolated is calculated in accordance with the direction detected to have the highest correlation from the total of 7 types of directions including the vertical direction and the 6 types of diagonal directions. The calculation procedure will be described below.

If the direction in which the pixel X to be interpolated has the highest correlation is the vertical direction, the interpolation direction of the pixel X to be interpolated is the vertical direction, and the pixel value of the pixel X is the upper and lower pixels D. And the average value of d is {(D + d) / 2}.

The direction in which the pixel X to be interpolated is detected to have the highest correlation is any one of the upward-sloping diagonal directions (diagonal direction Ag, diagonal direction Bf, or diagonal direction C).
In the case of e), or in the case of any one of the upward-sloping diagonal directions (diagonal direction Ec, diagonal direction Fb, or diagonal direction Ga), the direction detected as having the highest correlation with reference pixels X1 and X2 is referred to. Thus, the interpolation direction of the pixel X to be interpolated is determined.

That is, the direction detected as having the highest correlation with respect to the pixel X to be interpolated is one of the diagonally upward left directions (oblique direction Ag, oblique direction Bf, or oblique direction C).
In the case of e), the direction in which the highest correlation is detected for the reference pixel X1 is opposite to the direction in which the highest correlation is detected for the pixel X to be interpolated. LC, diagonal direction MB, or diagonal direction NA), or the direction detected as having the highest correlation for the reference pixel X2 is opposite to the direction detected for the pixel X to be interpolated as the highest correlation. Any one of the upward slanting directions (diagonal direction ej, diagonal direction fi,
Alternatively, in the diagonal direction gh), the interpolation direction of the pixel X to be interpolated is the vertical direction, and the pixel value of the pixel X to be interpolated is the average value {(D + d) / 2} of the upper and lower pixels D and d.

The direction in which the pixel X to be interpolated is detected to have the highest correlation is any one of the upward-sloping diagonal directions (diagonal direction Ag, diagonal direction Bf, or diagonal direction C).
In the case of e), the direction in which the highest correlation is detected for the reference pixel X1 is the same as the direction in which the highest correlation is detected for the pixel X to be interpolated. , Diagonal direction IF, or diagonal direction JE), or the vertical direction and reference pixel X
The direction in which the highest correlation is detected for 2 is the same as the direction in which the highest correlation is detected for the pixel X to be interpolated, which is one of the diagonal directions to the left (oblique direction an, oblique direction bm, or oblique direction). cl), or if it is the vertical direction, the interpolation direction of the pixel X to be interpolated is the diagonal direction in which it is detected that the pixel X to be interpolated has the highest correlation. If the interpolation direction is the diagonal direction Ag, the pixel value of the pixel X to be interpolated is the average value {(A + g) / 2} of the pixels A and g, and if the interpolation direction is the diagonal direction Bf, the pixel X to be interpolated. Is the average value {(B + f) / 2} of the pixels B and f, and if the interpolation direction is the oblique direction Ce, the pixel value of the pixel X to be interpolated is the average value {(C + e) of the pixels C and e. / 2}.

Similarly, in the case where the direction having the highest correlation with respect to the pixel X to be interpolated is any of the upward-sloping oblique directions (the oblique direction Ec, the oblique direction Fb, or the oblique direction Ga), the correlation with the reference pixel X1 is the highest. Is higher than the direction detected to have the highest correlation for the pixel X to be interpolated, whichever diagonal direction to the left (oblique direction HG, oblique direction IF, or oblique direction JE).
Or the direction in which the correlation is detected highest for the reference pixel X2 is opposite to the direction in which the correlation is detected highest for the pixel X to be interpolated. , Diagonal direction bm or diagonal direction cl), the interpolation direction of the pixel X to be interpolated is the vertical direction, and the pixel value of the pixel X to be interpolated is the average value of upper and lower pixels D and d {(D + d) / 2}.

The direction in which the pixel X to be interpolated is detected to have the highest correlation is any one of the upward-sloping diagonal directions (diagonal direction Ec, diagonal direction Fb, or diagonal direction G).
In the case of a), the direction in which the highest correlation is detected for the reference pixel X1 is the same as the direction in which the highest correlation is detected for the pixel X to be interpolated. , Diagonal direction MB, or diagonal direction NA), or the vertical direction and reference pixel X
The direction in which the highest correlation is detected for 2 is the same as the direction in which the highest correlation is detected for the pixel X to be interpolated, which is one of the upward-sloping diagonal directions (diagonal direction ej, diagonal direction fi, or diagonal direction). gh), or in the vertical direction, the interpolation direction of the pixel X to be interpolated is the diagonal direction in which the pixel X to be interpolated has the highest correlation. If the interpolation direction is the diagonal direction Ec, the pixel value of the pixel X to be interpolated is the average value {(E + c) / 2} of the pixels E and c, and if the interpolation direction is the diagonal direction Fb, the pixel X to be interpolated. Is the average value {(F + b) / 2} of the pixels F and b, and if the interpolation direction is the diagonal direction Ga, the pixel value of the pixel X to be interpolated is the average value {(G + a) of the pixels G and a. / 2}.

FIGS. 4 and 5 are diagrams for explaining highly accurate interpolation realized by the diagonal correlation adaptive scanning line interpolation method of the third embodiment. FIG. 4 is a diagram for explaining that an input image of an oblique line which is jagged in the prior art can be smoothly expressed, and FIG. 5 shows a part of a cross in which an oblique correlation is erroneously detected in the prior art. It is a figure explaining that an input image can be ideally interpolated without erroneous detection.

In FIGS. 4 and 5, the scan line (n-
1), scan line (n), scan line (n + 1), scan line (n +
2) is an actual scanning line. Also, A, B, C, D, E,
F and G are adjacent pixels on the actual scanning line (n), a, b,
c, d, e, f, g are adjacent pixels on the real scanning line (n + 1), and H, I, J, K, L, M, N are real scanning lines (n-).
1) Adjacent pixels on the top, h, i, j, k, l, m, n are adjoining pixels on the actual scanning line (n + 2), X is a pixel to be interpolated, and X1 and X2 are reference pixels. The pixel D is a pixel directly above the pixel X to be interpolated, the reference pixel X1 is a pixel directly above the pixel D, and the pixel K is a pixel directly above the reference pixel X1. The pixel d is a pixel directly below the pixel X to be interpolated, the reference pixel X2 is a pixel directly below the pixel d, and the pixel k is a pixel directly below the reference pixel X2.
Also, the white circles and black circles are pixels with different pixel values,
The pixel values of white circles and black circles are almost equal.

In the diagonal line image of FIG. 4, the direction detected as having the highest correlation in the pixel X to be interpolated by the interpolation method of the third embodiment is the diagonally upward right direction Ec. Further, the direction detected as having the highest correlation in the reference pixel X1 and the direction detected as having the highest correlation in the reference pixel X2 are both vertical directions. Since the direction with the highest correlation in the reference pixels X1 and X2 is the vertical direction, the interpolation direction of the pixel X to be interpolated is the upward-sloping diagonal direction Ec with the highest correlation, and the pixel value of the pixel X to be interpolated is , The average value of the pixels E and e in the diagonally upward right direction Ec {(E + c) /
2}. The interpolated scan lines generated in this way can smoothly represent diagonal lines.

Further, in the partial image of the cross in FIG.
The direction detected as having the highest correlation in the pixel X to be interpolated by the interpolation method of the third embodiment is the upward-sloping diagonal direction Fb. Further, the direction detected as having the highest correlation in the reference pixel X1 is the vertical direction, and the direction detected as having the highest correlation in the reference pixel X2 is the diagonally upward left direction bm. Since the direction having the highest correlation in the reference pixel X2 is the diagonal direction opposite to the direction having the highest correlation in the pixel X to be interpolated, the interpolation direction of the pixel X to be interpolated is the vertical direction and should be interpolated. The pixel value of the pixel X is the average value {(D + d) / 2} of the pixels D and d in the vertical direction. Therefore, it is possible to prevent the erroneous detection that has occurred in the conventional technique, and the interpolated scanning line thus generated can ideally interpolate the cross image as shown in FIG.

As described above, according to the third embodiment, two pixels above and below the pixel X to be interpolated, that is, seven pixels on a total of four actual scanning lines are used, and the vertical direction and the six kinds of diagonal directions are used. By detecting the direction having the highest correlation for the pixel X to be interpolated from the total of seven types of directions, it is possible to prevent erroneous detection of diagonal correlation, suppress image quality deterioration, and perform highly accurate interpolation.

In the third embodiment, the reference pixel X
The correlation for 1 and X2 was detected using 7 pixels on the actual scanning line above and below the reference pixel, for a total of 14 pixels, but for the reference pixels X1 and X2, the direction with the highest correlation is
Since it suffices to know whether it is the vertical direction, the diagonal direction to the left, or the diagonal direction to the right, the reference pixels X1, X
It is possible to reduce the number of pixels on the upper and lower actual scanning lines for detecting the correlation for 2 to 3 at the minimum. For example, as in the first embodiment, the actual scanning line (n-1)
The pixels I, J, K, L, and M above and the pixels B, C, D, E, and F on the actual scanning line (n) may be used to find the correlation for the reference pixel X1. By doing so, the correlation for the reference pixel can be obtained more easily.

In the third embodiment, the correlation for the pixel X to be interpolated is detected using seven pixels on the actual scanning line above and below the pixel X to be interpolated, that is, a total of 14 pixels. According to the invention, it is possible to reduce the number of pixels on the upper and lower actual scanning lines for detecting the correlation for the pixel X to be interpolated to at least three. For example, if it is sufficient to set the interpolation direction to the direction having the highest correlation of the three directions in total, that is, the vertical direction and the two diagonal directions (upward leftward diagonal direction and upward rightward diagonal direction), the actual scanning line (n) Pixels C, D, and E on the top and pixels c, d, and
The correlation for the pixel X to be interpolated may be obtained using e and e. By doing so, the correlation for the pixel to be interpolated can be obtained more easily.

[0110]

As described above, according to the scanning line interpolation method according to the first aspect of the present invention, the scanning line conversion device according to the sixth and eighth aspects, and the image display device according to the ninth aspect,
From the vertical direction and multiple diagonal directions, the vertical direction is given the highest priority, the diagonal directions that are as close to the vertical direction as possible are sequentially prioritized, and the direction with the highest correlation is detected to reduce false detection of diagonal correlation. There is an effect that highly accurate interpolation can be performed.

According to the scanning line interpolation method of the second aspect,
Seven pixels on each real scanning line are used for the first and second real scanning lines, and there are a total of seven directions in the vertical direction, three kinds of oblique directions rising to the right and three kinds of oblique directions rising to the left. Therefore, the vertical direction is given the highest priority, the diagonal directions that are as close to the vertical direction as possible are sequentially prioritized, and the direction with the highest correlation is detected. There is.

According to the scanning line interpolation method of the third aspect, the scanning line conversion device of the seventh and eighth aspects, and the image display device of the ninth aspect, at least two pixels above and below the pixel to be interpolated, a total of four pixels. By detecting the direction with the highest correlation for the pixel to be interpolated from among the vertical direction and multiple diagonal directions using the pixels on the actual scanning line, false detection of diagonal correlation is prevented and image quality deterioration is suppressed. The effect is that highly accurate interpolation can be performed.

According to the scanning line interpolation method of claim 4,
Detecting the direction with the highest correlation for the pixel to be interpolated and each of the first and second reference pixels, and referring to these detected directions to determine the direction with the highest correlation for the pixel to be interpolated. Thus, there is an effect that false detection of diagonal correlation can be prevented, image quality deterioration can be suppressed, and highly accurate interpolation can be performed.

According to the scanning line interpolation method of claim 5,
For the first to fourth actual scanning lines, seven pixels on each actual scanning line are used, and there are a total of seven types in the vertical direction, three types of diagonal directions to the right and three types of diagonal directions to the left. By detecting the direction with the highest correlation from the directions, it is possible to prevent erroneous detection of diagonal correlation, suppress image quality deterioration, and perform highly accurate interpolation.

[Brief description of drawings]

FIG. 1 is a diagram illustrating an oblique correlation adaptive scanning line interpolation method according to a third embodiment of the present invention.

FIG. 2 is a diagram illustrating an oblique correlation adaptive scanning line interpolation method according to the first embodiment of the present invention.

FIG. 3 is a diagram illustrating an oblique correlation adaptive scanning line interpolation method according to a second embodiment of the present invention.

FIG. 4 is a diagram for explaining highly accurate interpolation realized by the diagonal correlation adaptive scanning line interpolation method according to the third embodiment of the present invention, which is an input image of a diagonal line which is jagged in the prior art. It is a figure explaining that can be expressed smoothly.

FIG. 5 is a diagram for explaining highly accurate interpolation realized by the diagonal correlation adaptive scanning line interpolation method according to the third embodiment of the present invention, in which diagonal correlation was erroneously detected in the conventional technique. It is a figure explaining that the input image showing a part of crosses can be ideally interpolated without erroneous detection.

FIG. 6 is a diagram for explaining a conventional vertical interpolation type scanning line interpolation method.

FIG. 7 is a diagram illustrating a conventional diagonal correlation adaptive scanning line interpolation method.

FIG. 8 is a diagram illustrating a conventional diagonal correlation adaptive scanning line interpolation method.

FIG. 9 is a diagram illustrating a problem of a conventional diagonal correlation adaptive scanning line interpolation method, and is a diagram of an input image representing a part of a cross.

10 is a diagram illustrating a problem of a conventional diagonal correlation adaptive scanning line interpolation method, which is a diagram of an interpolated image of the image of FIG. 9 generated by the conventional diagonal correlation adaptive scanning line interpolation method. is there.

11 is a diagram for explaining the problems of the conventional diagonal correlation adaptive scanning line interpolation method, and is a diagram of an ideal interpolated image of the image of FIG. 9;

FIG. 12 is a block configuration diagram of a scanning line conversion device.

FIG. 13 is a block configuration diagram of a scanning line conversion device to which a scanning line interpolation method according to a third embodiment of the present invention is applied.

[Explanation of symbols]

1 input terminal, 2,3 field memory, 4 motion detection circuit, 5,11,12 line memory, 6
Double speed conversion circuit, 7 Moving image part interpolation scanning line generation filter, 8 Still image part interpolation scanning line generation filter, 9
Selector circuit, 10 output terminals.

Claims (9)

[Claims] 1. A diagonal-correlation-adaptive scanning line interpolation method, at least a first actual scanning line above a pixel to be interpolated,
By using the pixels on a total of two actual scanning lines of the second actual scanning line below the pixel to be interpolated, the vertical direction from the vertical direction and the plurality of diagonal directions centering on the pixel to be interpolated Is the highest priority, the diagonal direction as close as possible to the vertical direction is sequentially prioritized, and the correlation detection step of detecting the direction having the highest correlation among the pixels to be interpolated; and the correlation detection step on the first and second actual scanning lines. And a step of setting an average value of two pixels in a direction having a high correlation as a pixel value of the pixel to be interpolated, the scanning line interpolation method. 2. The scanning line interpolation method according to claim 1, wherein the correlation detecting step uses at least seven pixels on each of the first and second actual scanning lines in the vertical direction. In addition, a scanning line interpolation method is characterized in that the direction having the highest correlation is detected from a total of seven types of directions, namely, three types of diagonal directions to the right and three types of diagonal directions to the left. 3. A diagonal-correlation adaptive scanning line interpolation method, comprising: at least a first actual scanning line above a pixel to be interpolated;
The second real scan line below the pixel to be interpolated;
Pixels on a total of four real scanning lines including the third real scanning line above the real scanning line and the fourth real scanning line below the second real scanning line A correlation detecting step of detecting a direction having the highest correlation with respect to the pixel to be interpolated from the vertical direction and a plurality of diagonal directions centering on the center, and the highest correlation on the first and second actual scanning lines. And a step of setting an average value of two pixels in a direction as a pixel value of the pixel to be interpolated, the scanning line interpolation method. 4. The scanning line interpolation method according to claim 3, wherein the correlation detection step uses at least pixels on the first and second actual scanning lines in a vertical direction centered on the pixel to be interpolated. And a direction having the highest correlation with respect to the pixel to be interpolated from among a plurality of diagonal directions, and using at least the pixels on the first and third actual scanning lines, the first actual scanning line and the above The direction having the highest correlation with respect to the first reference pixel is detected from the vertical direction and the plurality of oblique directions centering on the first reference pixel between the third actual scanning lines, and at least the second and Using the pixels on the fourth actual scan line, from the vertical direction and a plurality of diagonal directions centered on the second reference pixel between the second actual scan line and the fourth actual scan line, About the second reference pixel The direction having a high correlation is detected, and the direction to which the pixel to be interpolated and the direction to which the pixel to be interpolated is detected are determined to have the highest correlation with respect to the pixel to be interpolated. A scanning line interpolation method characterized by determining a high direction. 5. The scanning line interpolation method according to claim 3, wherein the correlation detection step is at least the first to fourth steps.
Up to 7 actual scanning lines, 7 pixels on each actual scanning line are used, and the most correlation is obtained from the total of 7 vertical directions, 3 upward diagonal directions and 3 upward diagonal directions. A scanning line interpolation method, which is characterized by detecting a direction having a high value. 6. At least a first above the pixel to be interpolated
, And a second real scanning line below the pixel to be interpolated, a total of two pixels on the real scanning line in the vertical direction and a plurality of diagonal directions centered on the pixel to be interpolated. From among the above, the vertical direction is given the highest priority, the diagonal directions which are as close to the vertical direction as possible are sequentially given priority, and the correlation detection means for detecting the direction with the highest correlation for the pixel to be interpolated; A scanning line conversion device comprising: an interpolating scanning line generating means for generating an interpolating scanning line by using an average value of two pixels in the direction having the highest correlation on the actual scanning line as a pixel value of the pixel to be interpolated. . 7. At least a first above pixel to be interpolated
Real scan line, a second real scan line below the pixel to be interpolated, a third real scan line above the first real scan line, and a second real scan line below the second real scan line. Using the pixels on a total of four real scanning lines of the fourth real scanning line, the correlation is highest for the pixel to be interpolated from the vertical direction and a plurality of diagonal directions centering on the pixel to be interpolated. Correlation detection means for detecting a direction, and interpolation for generating an interpolated scan line by using an average value of two pixels in the direction having the highest correlation on the first and second actual scan lines as a pixel value of the pixel to be interpolated. A scanning line conversion device comprising: scanning line generation means. 8. The scanning line conversion device according to claim 6 or 7, wherein the interpolation scanning line generating means is a moving image interpolation scanning line generating means, and further, motion detection for detecting a motion of an input video signal. Means, a still image interpolation scanning line generation means, and based on the result detected by the motion detection means,
A scanning line conversion apparatus comprising: a selection unit that selects one of a signal output from the moving image interpolation scanning line generation unit and a signal output from the still image interpolation scanning line generation unit. 9. An image display device, comprising the scanning line conversion device according to claim 6.