JPH06153169A - Progressive scan conversion method - Google Patents

Abstract

(57) [Summary] [Object] To provide a progressive scan conversion method with improved conversion image quality for moving images. When converting an interlaced-scanned television video signal into a sequentially-scanned video signal, pixels existing on and below each scanning line to be interpolated for progressive scanning conversion, and to be interpolated From among a plurality of two pixel groups sandwiching the pixel to be interpolated corresponding to a plurality of diagonal edges including, one or a plurality of groups having a small difference between the two pixels are identified, and the group is identified. The sum of the two pixels of the set is used as the interpolation output of the pixel to be interpolated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal scanning conversion method, and more particularly to a conversion method from 2: 1 interlaced scanning to progressive scanning.

[0002]

BACKGROUND ART Interlaced scanning is a method of transmitting a television signal with a required video signal bandwidth that is half that of progressive scanning by utilizing the property that human vision is less likely to perceive small-area flicker as compared to large-area flicker. is there. However, when interlaced scanning is performed, it is possible to halve the image band without causing large-area flicker, but when high image quality is required, it has been pointed out that deterioration in image quality cannot be ignored for sequential scanning. There is. From the viewpoint of sampling television scanning, what is the difference between interlaced scanning and progressive scanning, what is the case when converting an interlaced scanning video signal of a general television signal into a high-quality progressive scanning? Reference (1), Nishizawa, Tanaka: “Image quality improvement by progressive scan conversion”, NHK STRL, Vol.27, No.12, pp.27-32 ( 1984).

In addition, a motion compensation technology is used for conversion from interlaced scanning to progressive scanning to achieve high image quality.
Another application which becomes the applicant of this application, Reference (2), Japanese Patent Application No. 4-28
There is a 3120 “motion-compensated progressive scan conversion device”.

[0004]

According to the conventional progressive scan conversion technique, the conversion image quality for still images is good, but interlace interference is not removed for moving images, and the spatial frequency response in the vertical direction is about 1 or less. There is a drawback that the converted image quality is not very good because it drops to / 2. In particular, when the diagonal edge image moves, the edge becomes jagged depending on the moving speed, and the image is very unsightly.

According to the reference (2), a good converted image quality is obtained even for a moving image. However, at the present time, the motion vector detection circuit and motion compensation memory required to implement the technique disclosed in the reference (2) are expensive, and more economical means are also required. Therefore, it is an object of the present invention to provide a progressive scan conversion method capable of solving the above-mentioned drawbacks and economically realizing a good progressive scan conversion image for moving images, at least moving diagonal edges. is there.

[0006]

To achieve this object, the progressive scan conversion method of the present invention interpolates for progressive scan conversion in converting an interlaced scanned television video signal into a progressively scanned video signal. Among a set of two pixels sandwiching the pixel to be interpolated, which correspond to a plurality of diagonal edges including pixels to be interpolated and which are present on the scanning lines above and below the pixel to be interpolated, respectively. One or a plurality of sets having a small difference between two pixels are specified, and the sum of two pixels of the specified set is used as an interpolation output of the pixel to be interpolated.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the arrangement of a progressive scan conversion apparatus which is an embodiment of the method of the present invention. The input signal of the device is 525 scanning lines / field frequency 60Hz /
It is a television image signal with an interlace ratio of 2: 1 and the output signal is an image signal of 525 lines / frame frequency 60 Hz / sequential scanning 1: 1. Both the input and output signals are digital signals, the sampling frequency of the input signal is μs (MHz), and the sampling frequency of the output signal is 2 μs (MHz).

In FIG. 1, the input signal is 2D first.
(2 pixels) is input to the delay device 1 and delayed by 2 pixels. This is for phase matching with the intra-field interpolation circuit 42 described later. The output of the delay device 1 is the motion detection circuit 2
Is input to each pixel to determine whether the pixel is a still image pixel or a moving image pixel. As this circuit, various conventionally proposed motion detection circuits can be used as they are.

The motion detection circuit 2 outputs the mixing ratio k of the static input and the dynamic input in the mixing circuit 4. k is 0, for example
≦ k ≦ 1, k = 0 for still pixels and k for moving pixels
= 1. The output of the delay device 1 is also input to the 263H delay device (H is one line in the field) 3. The output of the delay 3 is the interpolated signal from the previous field for the stationary pixel and is applied to the stationary input of the mixing circuit 4. The output of the intra-field interpolation circuit 42 described later is added as a dynamic input to the mixing circuit 4. The mixing circuit 4 mixes the two inputs according to the above-mentioned k and outputs, for example, (1-k) × (stationary input) + k × (moving input).

The output of the mixing circuit 4 is an interpolation signal for interpolating scanning lines between scanning lines by interlaced scanning. The output signal of the delay device 1 and the interpolation signal are combined by the time axis conversion circuit 5 to be converted into a sequential scanning signal, which becomes an output signal of this conversion device. The contents and operation of the conversion circuit 5 are described in detail in Reference (2), for example.

On the other hand, the input signal of the converter is an H delay device (1
It is also input to the line delay) 6 and the D delay device (1 pixel delay) 7. When the input of the delay device 7 is the signal j, the signal j becomes the signals i, h, g, f through the D delay devices 7, 8, 9, 10. When the output of the delay device 6 is the signal e, the signal e is D
Signals d, c, b, and a are sequentially passed through the delay units 11, 12, 13, and 14. The subtracters 15 to 19 respectively output the signals S ₁ = (ja) and S ₂ = (ib)
, S ₃ = (hc), S ₄ = (gd), S ₅ = (fe) are output. Similarly, the adders 20 to 24 have signals A ₁ = (j + a), A ₂ = (i + b),
Outputs A ₃ = (h + c), A ₄ = (g + d), A ₅ = (f + e).

The signals S _{1 to} S ₅ are quantizers having the same characteristics.
The numbers of bits in the amplitude direction are appropriately compressed by 37 to 41 to become signals T _{1 to} T ₅ . Signals T _{1 to} T ₅ are all decision circuits
Entered in 25. The judgment circuit 25 can be realized by a ROM, for example. If the signal T ₁ through T ₅ far is if compressed in the 3 bits, the decision circuit 25 address input is 15 bits, i.e. 32
It can be realized with a k-word ROM. The judgment algorithm can be set freely if it is implemented in ROM. Some examples are given later. The output of the determination circuit 25, which is the determination result, is the values of the coefficients α _{1 to} α ₅ , which are α _{1 to} α ₅ coefficient units 31 to
Entered in 35. On the other hand, the signals A _{1 to} A ₅ are input to the α _{1 to} α ₅ coefficient units 31 to 35 after passing through the 1/2 coefficient units 26 to 30, respectively. The outputs of the coefficient units 31 to 35 are all added in the adder 36.

The delay device 6 to the adder 36 as a whole constitute the intra-field interpolation circuit 42 according to the method of the present invention, and the output of the adder 36 is the output of the intra-field interpolation circuit 42 and the dynamic input of the mixer 4. Added to. Next, the operation of the device of the present invention will be described with reference to FIG. FIG. 2 is a two-dimensional view of an image in the horizontal (x) -vertical (y) direction. The horizontal solid lines are scan lines of the current field by interlaced scanning, and the horizontal dotted lines are scan lines to be interpolated for progressive scan conversion. Pixels corresponding to the signals a to j shown in FIG. 1 are shown by white circles in the figure. The horizontal distance between pixels is 1 / μs, and the distance between scanning lines is Ph / 525 (Ph: screen height).

Consider now the pixel h (double circle) and create the interpolated pixel x for it. The image is assumed to be a uniform diagonal edge having the inclination as shown in the figure. That is, in FIG. 2, the upper left region (pixel a,
b, c, d, x, f, and g) are white, and the lower right region (including pixels e, h, i, and j) is black. If perform spatially symmetric interpolation with respect to the interpolation point x, it is clear from FIG this case g + d / 2 (= A 4/2) that the x is the best. In conventional intra-field interpolation, h + c / 2 is always interpolated, so in the case shown in the figure, the interpolated pixel is at an intermediate level between white and black, and an interpolation error occurs. This is one of the reasons why the image quality of the conventional intra-field interpolation is not good. According to the method of the present invention, such a phenomenon does not occur and a slanted edge can be completely reproduced.
However, since the slanted edges may have various inclinations, it is necessary to prepare an interpolation signal corresponding to each inclination and determine which of them is selected from the image content. Figure 1 in the illustrated apparatus 5 of the interpolation signal A ₁ to A ₅
The types are prepared and the selection is determined by the difference signals S _{1 to} S ₅ corresponding to them.

An example of determination will be described with reference to FIG. The figure is a horizontal (μ) -vertical (ν) two-dimensional spatial frequency spectrum, and the unit of ν is cph (cycles per height). The spectrum of an oblique edge (hereinafter abbreviated as gd edge) having the same slope as that of the pixel gd shown in FIG.
Then, it becomes a spectrum on a straight line extending from the origin to the upper left.
Also, since the image is interlaced scan (μ, ν)
= (0,525 / 2) is the carrier for sampling by interlaced scanning, and the folded spectrum (dotted line) by interlaced extends from there to the lower right. spatial frequency response for the input signal of the differential signal S ₄ corresponding to the g-d edges, the stop band is shown as shaded areas in FIG. The area not shaded is the pass area.
As is clear from the figure, since the spectrum of the gd edge is entirely included in this stop band, S ₄ = 0 when the image is the gd edge. In contrast, the image is ef
For edge signal remains in the pass band of S ₄ for the Superutoru as shown in dashed line in FIG, | a> 0 | S _4. It is possible to determine the slope of the thus image edges by the differential signal S ₁ to S _5.

These five differential signals S _{1 to} S ₅ (in FIG. 1, T _{1 obtained} by bit-compressing S _{1 to} S ₅ for circuit scale reduction)
There are many concrete algorithms (contents of the determination circuit 25 in FIG. 1) for selecting one gradient from T ₅ to T ₅ ), but an example is shown below.

Judgment Algorithm: Example 1

[Equation 1] In this case, the coefficient units 31 to 35 and the adder 36 in FIG. 1 are equivalent to one switch. Judgment algorithm: Example 2

[Equation 2] In this example, when the edge slope is an intermediate slope between the slopes assumed by the prepared interpolation filter, the output is obtained by linear interpolation from the two interpolation filters sandwiching it.

Now, the interpolation filter A ₁
When properly selected in any of to A _5, the folded spectrum by interlacing can be completely removed by interpolation. As an example, the impulse response and spatial frequency response of the interpolation filter for the gd edge are shown in Fig. 4 (a),
Shown in (b). In Fig. 4 (b), the shaded area is the filter stop area, and it is clear that the interlaced folds have been completely removed.

The method of the present invention is not limited to the embodiment, but may be, for example, 6
It can also be used for 25 systems and HDTV. It is also possible to adapt to an image with a wider range of inclination by changing or increasing the delay amount of each unit, the number of delays, and the circuits associated therewith.

[0020]

According to the method of the present invention, by using the interpolation filter in the diagonal direction adapted to the contents of the image, the interlace interference can be removed without damaging the edge information for the moving image, at least for the moving diagonal edge, and the interlace The conversion image quality of the progressive scanning conversion can be improved greatly and economically.

[Brief description of drawings]

FIG. 1 is a block diagram of an embodiment of an apparatus according to the method of the present invention.

FIG. 2 is a diagram for explaining the operation of the apparatus shown in FIG.

FIG. 3 is a diagram showing a horizontal-vertical spectrum of an oblique edge.

FIG. 4 is a diagram for explaining an example of a diagonal interpolation filter.

[Explanation of symbols]

1 2 pixel delay device 2 motion detection circuit 3 263 line delay device 4 mixing circuit 5 time axis conversion circuit 6 1 line delay device 7 to 14 1 pixel delay device 15 to 19 subtractor 20 to 24, 36 adder 25 determination circuit 26 ˜30 1/2 coefficient unit 31 to 35 α _i coefficient unit 37 to 41 quantizer 42 intra-field interpolation circuit

Claims (1)

[Claims] 1. When converting an interlaced-scanned television video signal into a progressively-scanned video signal, the interlaced-scanned television video signal is present on and above each scan line above and below a pixel to be interpolated for progressive scan conversion. From among a plurality of two pixel groups sandwiching the pixel to be interpolated corresponding to a plurality of diagonal edges including a power pixel, one or a plurality of groups having a small difference between the two pixels are specified,
A progressive scan conversion method, wherein the sum of two pixels of the specified set is used as an interpolation output of the pixel to be interpolated.