JPH0595537A - Image data processing method - Google Patents

Abstract

(57) [Summary] (Modified) [Purpose] To realize image data compression / decompression with a simple circuit configuration with little distortion. [Structure] When performing compression / expansion processing of an image data string, the ratio of the number of data in the input data string to the number of data in the output data string is set to a simple integer ratio.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data processing system for compressing / decompressing an image signal, and more particularly to an EDTV.
In the (Extended Definition TV) system and the like, the present invention relates to an image data processing system used when compressing / decompressing a signal data sequence such as high resolution information or side panel information inserted in a gap of an NTSC video signal.

[0002]

2. Description of the Related Art EDT that achieves high resolution and wide aspect while maintaining compatibility with the current NTSC system
Various V methods have been proposed. In this EDTV system, as a means for achieving high resolution and wide aspect, a signal data string such as high resolution information and side panel information is provided, for example, in a blowhole or horizontal / vertical overscan,
A method is used in which the image carrier is inserted into a gap such as an orthogonal component and transmitted. For the signals inserted into these gaps, basically, it is a characteristic of the individual system as to what kind of component and form it takes. Here, a data processing process called compression / expansion of an insertion signal data sequence is generally known as one data processing method used when inserting and transmitting image data in this gap and reproducing it. There is.

7 to 11 are explanatory views of a conventional example in the case of compressing / expanding an insertion signal data string of side panel information or the like. As shown in FIG. 7, the wide-aspect screen 100 in the EDTV system is displayed on the center panel section CP.
Side panel right end SPR and side panel left end SP
When divided into three regions of L, consider a case where, for example, the right end SPR and the left end SPL of the side panel section are compressed and inserted into the overscan region HOS of the screen 101 having an aspect ratio of 4: 3 and transmitted.

Generally, the left and right side panel parts SPR and SPL in the side panel system of the EDTV are about 30% of the area of the center panel part CP, and here it is assumed that sampling is performed at a sampling frequency of 4 fsc. Then, it is assumed that the left end and the right end of the side panel portion each correspond to an area of 100 pixels (about 7 μsec).

First, a low-pass luminance component of the image signal of the side panel portion is low-pass filtered (hereinafter referred to as "LPF").
It is considered that the band is limited by, pre-processing is performed so that the band does not exceed the band after the signal is compressed, and then the signal is compressed into 23 pixels (about 1.6 μsec) as shown in FIG.

In this case, the left and right side panel parts SPR,
SPL 100 pixel sampling point (FS1 = 4fs
c) is thinned out to 23 pixels (FS2 = 4fsc × 23/100). Therefore, the compression ratio of the signal is 100: 2.
3, that is, 4.3478, and since 23 is a prime number, it is not a simple integer ratio.

Therefore, when reproducing the image signal of the side panel section compressed to the above 23 pixels to the original video signal of 100 pixels, in this example, all the new sampling points are calculated by the interpolation formula of FIG. 9, for example. It will have to be calculated and obtained. Although the difference from the ideal value and the scale of the hardware may differ depending on the form of the interpolation function, for example, as the interpolation function, as shown in FIG. Ideal interpolation data can be obtained by using.

However, the sampling frequency FS2
To create a new sample value by sampling with FS
It is necessary to use a large number of sampled values of 1 in a pseudo manner, and the hardware becomes extremely complicated, which is not practical.

On the other hand, when linear interpolation is used as the interpolation function as shown in FIG. 10B, the load on the hardware is reduced, but due to its nature, the response of the sampling frequency characteristic from the low range is low. Since the response is lowered and the response does not become zero even in a high frequency range, when sampling is performed again at the sampling frequency FS2, there may be a case where the aliasing distortion is repeatedly generated while being attenuated with time as shown in FIG. ..

[0010]

When the ratio of the sampling frequency FS1 to the resampling frequency FS2 is selected to be a halfway value as in the above-mentioned conventional apparatus, interpolation data for creating a new sample value can be created. When a sampling function is used, an ideal one with less distortion can be obtained, but there is a problem that a large-scale hardware is required and the circuit configuration becomes complicated and expensive.

Further, when linear interpolation is used as the interpolation function, the circuit configuration becomes simple but folding distortion occurs,
In particular, when the original data is reproduced by using compression and expansion as a pair, there is a problem that a large distortion is generated as shown in FIG.

That is, in FIG. 11, the shaded portions 103, 10
As indicated by reference numerals 4 and 105, distortion as a filter occurs in the frequency characteristics at the time of compression and expansion and the overall characteristics, and unnecessary spectra 106, 107, and 108 are generated due to aliasing distortion that repeatedly occurs while decaying with time.

As described above, any of the interpolation processes has a disadvantage. This is because the compression ratio / expansion ratio was not selected as a simple integer ratio.

[0014]

In order to solve the above problems, the present invention provides an image signal obtained by inserting a first image data sequence obtained by compressing a first image signal into a gap between second image signals. In the image data processing method in which the first image data string is extracted, the first image data string is decompressed, and the original first image signal is reproduced, The compression / expansion ratio is selected so that the ratio of the sampling number when performing the compression process and the sampling number when performing the expansion process on the first image data string is a simple integer ratio.

[0015]

Since the present invention is configured as described above, the first image signal extracted from the second image signal is used as the basis for the sampling number of the first image signal to be inserted into the second image signal. The ratio of the number of samplings when expanding to the first image signal can be a simple integer value.

That is, for example, assuming that the video signal of the side panel section is in the range of about 100 pixels (about 7 μsec) on the left and right, and the picture signal is inserted into the horizontal overscan section after being compressed, the horizontal overscan width is about 0. .7
Assuming ˜2 μsec, the integer ratio of compression / expansion that can be realized in this case is 7 / 0.7 (10) as shown in FIG.
Up to 7/2 (3.5) integer ratios 10, 9, 8, 7,
6, 5, 4 can be cited, and any integer ratio can be taken.

When the same thing is to be realized by using the blow-out hole, as shown in FIG. 6, about 100 pixels (about 7 μsec) of each of the left and right video signals of the side panel section are converted to the video signal of the center panel section. It extends to about 54 μsec and is inserted into the blow-out hole, but since the video signal of the side panel portion of about 7 μsec × 2 = 14 μsec in total is extended to about 54 μsec,
The integer ratio of compression / expansion is 54: 14 = 3.86, and 3 or 4 is selected as the realizable integer ratio in this case.

[0018]

1 to 6 are explanatory views of an embodiment of the present invention. In Fig. 1, the luminance low-frequency component (YSL) of the side panel part (100 pixels each on the left and right ... Approximately 7 μsec) of the wide aspect screen is compressed to 25 pixels (approximately 1.746 μsec) at both ends of the screen that becomes the overscan portion of the screen The case of overlapping is illustrated.

In this embodiment, 100 of the side panel portion is used.
The pixel sampling point (FS1 = 4fsc) is set to 25 pixels (FS2 = 4fsc × 25/10) in the overscan section.
It will be thinned out to 0). Here, the compression ratio described above is 100: 25, that is, 4, and the ratio of input / output data is a simple integer ratio. Therefore, in this example, as shown in FIG. 2, it is only necessary to thin out 1 data out of 4 input data by the compression circuit 1 and take out all the output data strings, and it is very easy to generate the interpolated data. It is not necessary to calculate using the formula of insertion.

On the other hand, the process of decompressing the data string compressed by the above method and restoring it is as shown in FIG. Figure 3
The compressed FS2 data string shown in (a) is shown in FIG.
As shown in (4), conversely to the compression processing, four data are obtained from one data by interpolation.

FIG. 4 shows a decompression circuit for reproducing the FS1 data string from the FS2 data string shown in FIG. 3 (a) by interpolation / interpolation processing, and is supplied from the input terminal 4 and shown in FIG. 3 (a). The data string is zero-inserted by the pseudo FS1 data string conversion circuit 2 and the pseudo FS1 as shown in FIG.
It is converted into a data string and guided to the interpolation processing circuit 3 in the next stage.

The interpolation processing circuit 3 interpolates the pseudo FS1 data string from the pseudo FS1 data string conversion circuit 2 by a predetermined interpolation function and outputs an FS1 data string as shown in FIG. 3 (d). It is output to the terminal 5, and in this embodiment, an LPF having a characteristic as shown in FIG. 3C is used as an interpolation function.

Therefore, in the decompression processing, the degree of interpolation error and the scale of hardware differ depending on what is selected as the interpolation function. For example, even if a sampling function is used as the interpolation function, the Sampling frequency FS1
Sampling frequency F to create a new sampled value by
It can be realized with an extremely simple configuration in which the sample value of S2 is artificially multiplied by four, which is very practical.

On the other hand, even when linear interpolation is used as the interpolation function, the true value is sampled once every four data. Therefore, for example, compression processing is followed by decompression processing to obtain image data. When performing reproduction, distortion (FIG. 11) due to re-sampling twice before and after as shown in the conventional example of FIG. 11 is reduced by sampling both true values. In this way, it is an optimum and effective means not only for simplifying the hardware but also for reducing distortion at the time of image reception.

In the above embodiment, in the EDTV side panel system, a video signal of 100 pixels on each of the left and right sides of the side panel is compressed to 25 pixels on each of the left and right overscan portions of the screen (compressed 4 times). Transmitted and then expanded again to the original 100 pixels (expanded 4 times)
However, the compression / expansion rate is not limited to four times, and any simple integer, for example, a one-digit integer may be used.

The application is not limited to the signal processing in the panel system of the EDTV shown in the embodiment, but can be widely applied to the signal processing in which compression and expansion are performed in pairs.

[0027]

Since the present invention is constructed as described above,
In the compression / decompression process of image data, not only the distortion of reproduced data in the interpolation process can be reduced, but also the hardware configuration at the time of the interpolation process can be simplified.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of an embodiment of signal compression in the present invention.

FIG. 2 is a block diagram showing a configuration of an embodiment at the time of signal compression in the present invention.

FIG. 3 is an explanatory view of an embodiment at the time of signal expansion in the present invention.

FIG. 4 is a block diagram showing the configuration of an embodiment at the time of signal expansion in the present invention.

FIG. 5 is an explanatory diagram of another embodiment at the time of signal compression in the present invention.

FIG. 6 is an explanatory diagram of another embodiment at the time of signal expansion in the present invention.

FIG. 7 is an explanatory diagram of a signal processing method in the center panel system.

FIG. 8 is an explanatory diagram of signal compression of a conventional example.

FIG. 9 is an explanatory diagram of interpolation processing at the time of signal expansion in the conventional example.

FIG. 10 is a frequency characteristic diagram of the interpolation method at the time of signal expansion in the conventional example.

FIG. 11 is a frequency characteristic diagram of aliasing distortion in the conventional linear interpolation method.

[Explanation of symbols]

 1 signal compression circuit 2 pseudo FS1 data string conversion circuit 3 interpolation processing circuit

Claims (1)

[Claims] 1. A first image data string is extracted from an image signal in which a first image data string obtained by compressing a first image signal is inserted into a gap between second image signals, and the first image data string is extracted. In the image data processing method in which the data string is expanded and the original first image signal is reproduced, the number of samplings when compressing the first image signal and the first number
The image data processing method is characterized in that the compression / expansion ratio is selected so that the ratio to the sampling number when the image data string is expanded is a simple integer ratio.