JPH04170185A - Picture processing method - Google Patents

Abstract

PURPOSE:To improve the picture quality of a display picture by applying inverse DCT arithmetic operation to a corrected matrix data so as to form a picture data, converting the picture data into an analog data thereby obtaining a video signal. CONSTITUTION:When a picture data results from a television signal, DCT (discrete cosine transformation) (2) is applied to the picture data and a correlation correction data acquired statistically in advance is multiplied with or added to a DCT output data (matrix data) being the result of the arithmetic operation to correct, e.g. a high frequency component. The corrected DCT output data is subjected to inverse DCT arithmetic operation (9) to obtain the television signal. Moreover, when the picture data results from a VTR, a tape hiss data acquired statistically in advance is divided by or subtracted from (3) the DCT output data before the correction of the high frequency component to reduce noise. Thus, since missing or an attenuated part of the television signal such as a high frequency component is reproduced based on the correlation of horizontal and vertical spatial frequency, the picture quality of the picture is improved.

Description

[Detailed Description of the Invention] [Industrial Application Field] This invention uses D C'I' (discrete cosine transform) operation for image processing of VTRs and TV devices (television),
The present invention relates to an image processing method that corrects the high frequency range of a video signal to improve the quality of a displayed image.

[Conventional example] Conventionally, in image processing for VTRs and TV equipment, when correcting the high frequency range of a video signal, the normal analog system uses a compensation circuit or adjustment circuit (for example, a peaking, contour enhancement circuit, etc.) to correct the image. It compensates for and adjusts the high frequencies of the signal.

[Problems to be Solved by the Invention] However, in the above image processing method, high frequency correction and emphasis adjustment of the video signal are performed on the spatial frequency in the horizontal direction, and only a limited number of spatial frequencies are used in the vertical direction. It is only done for. Therefore, in VTRs and TV devices, tape hiss and satellite line noise may cause a reduction in image quality, especially natural image quality.

This invention was made in view of the above-mentioned problems, and its purpose is to provide image processing that uses the correlation between horizontal and vertical spatial frequencies to correct high-frequency components and improve image quality. The purpose is to provide a method.

[Means for Solving the Problems] In order to achieve the above object, the present invention digitally converts a video signal into image data, and converts this image data into D.
CT calculation is performed, and the matrix theta of this DCT calculation result is multiplied or subtracted by data obtained statistically in advance,
The high-frequency components are corrected using the correlation between the spatial frequencies in the horizontal and vertical directions, and the corrected matrix data is subjected to an inverse DCT operation to obtain image data. This image data is converted into an analog signal and converted into a video signal. The summary is what was done.

[Operation] Since the above method is used, if the image data is from a television, DCT calculation is performed on the image data,
Correlation correction data obtained statistically in advance is multiplied or added to the DCT output data (matrix data) resulting from the calculation, thereby correcting, for example, high-frequency components. This corrected DCT output data is subjected to an inverse DCT operation to obtain a television signal.

In addition, if the image data is from a VTR, before correcting the high frequency component, divide or subtract tape hiss data statistically obtained in advance from the DCT output data,
Reduce noise.

Therefore, the correlation between the horizontal and vertical spatial frequencies results in more missing or attenuated portions of the television signal.

For example, since high-frequency components are reproduced, image quality can be improved.

[Example] Hereinafter, an example of the present invention will be described based on FIGS. 1 to 4.

In the image processing method of the present invention, instead of using DCT (discrete cosine transform) to compress image data, noise is removed and the original data is processed using statistical correlation data for horizontal and vertical spatial frequencies. High-frequency corrections are made based on predictions.

Therefore, as shown in FIG.
/D conversion unit 1 and converts the converted image data into a two-dimensional DO
DCT section 2 that performs T calculation and this DOT output data (8X
A tape hiss/noise pattern number circuit 3 detects the noise pattern by comparing the 8-pixel 71-lix data) with a statistically obtained pattern of tape hiss or noise data from a satellite line, and a first correction coefficient table unit 4 that outputs the coefficients obtained by the DCT; a first calculation unit 5 that divides or subtracts the coefficients from the DCT output data to remove noise caused by tape hiss/satellite line, etc.; A pattern match circuit 6 that compares the removed T) CT output data with a high-frequency component pattern statistically obtained in advance and detects the pattern, and a second correction coefficient table that outputs coefficients according to this pattern match. a second calculation unit 8 which multiplies or adds this coefficient to the noise-removed DCT output data to compensate for missing or attenuated high-frequency components of horizontal and vertical spatial frequencies; DCT output data with high-frequency components corrected is inverted D
An inverse DCT unit 9 performs CT calculations, and a D/D converter converts the image data obtained by the inverse DCT calculations into analog to obtain a video signal.
A conversion section 10 is provided.

Next, we will explain the image processing method applied to the above-mentioned image processing device for anabolic acid with reference to the schematic diagrams illustrating the DCT output in FIGS. 2 and 3, and the spatial frequency correlation characteristic diagram in FIG. 4. explain.

First, assuming that the video from a tape or satellite line is manually generated, the video signal is converted into digital data by the AID converter 1 and converted into image data. in this case,
As shown in FIGS. 2 and 3, the image data of one screen is a 64×60 block, and this one block has 8×8 pixels. Then, in the DCT calculation unit 2, DCT calculation is performed on the converted image data, and DCT output data (matrix data of 8×8 pixels)
is obtained, and in this matrix data, as shown in FIG. 3, horizontal spatial frequency components appear in the horizontal direction of the figure, and vertical spatial frequency components appear in the vertical direction.

Next, in the case of a VTR device, the tape hiss/noise pattern matching circuit 3 compares the specific pattern of the DCT output data with a tape hiss noise pattern statistically obtained in advance. The data of the noise pattern is as shown in D above.
It may be obtained as statistical data based on the characteristics of the arrangement of matrix data of CT calculation results. Further, in the case of a TV device (television), the noise pattern is due to line noise. Based on the comparison results, the first correction coefficient unit 4 obtains coefficients corresponding to tape hiss noise and line noise from the spatial frequency correlation shown in FIG. The matrix data of the DCT output is divided or subtracted.

Then, after the noise due to tape hiss (or satellite line) is removed, the pattern match circuit 6 compares the noise-removed DCT output data with a pattern statistically obtained in advance, and the result of this comparison is used as the second is output to the correction coefficient table section 7. Based on the comparison result, the second correction coefficient section 7 obtains a coefficient for correcting the data of the DC'r output from the spatial frequency correlation function shown in FIG. The DCT output data from which noise has been removed is multiplied or added.

Here, when the image data to be improved is subjected to DCT calculation, missing or attenuated parts are compensated for based on the comparison result of the DCT output data of this calculation result and the above statistical pattern. . That is, the fourth
As shown in the figure, the horizontal spatial frequency (Hf
) and the spatial frequency (Vf) in the vertical direction are correlated with each other, so if there is attenuation in the vertical direction, for example, the amount of attenuation can be known. Note that the above statistical pattern is based on the characteristics of the data pattern in high-quality video in advance.
CT calculation is performed, and statistics are taken from the result of this DCT calculation to obtain the above-mentioned statistical pattern for detecting pattern matching in the pattern matching circuit 6. Further, the pattern matching circuit 6 detects what statistical pattern the pattern of the DCT output data (matrix data) matches, but it is only necessary to detect a match in a limited portion of the matrix data.

Therefore, for example, if the high frequency range of a manually generated video signal is attenuated, the second calculation unit 8 corrects the high frequency range, and also corrects both the horizontal and vertical spatial frequencies. is possible.

Next, the above-mentioned high-frequency corrected DCT output data is converted to inverse DC
The T calculation section 9 performs an inverse DCT operation, and the inverse DCT output data is analog-converted in the D/A conversion section 10 to form a video signal.

Images based on video signals obtained in this way have their high-frequency components corrected, and the correction has also been made for the horizontal and vertical spatial frequencies, so that the conventional horizontal signals, which tend to be degraded, can be corrected. Since the vertical spatial frequency, which is less susceptible to degradation, is used, the image quality of the displayed image can be improved, rather than performing high-frequency correction using the vertical spatial frequency.

[Effects of the Invention] As explained above, in the image processing method of the present invention, a human video signal is digitally converted into image data, and this image data is subjected to a DC"I" operation to be converted into matrix data (DCT output data). ), this matrix data is corrected with data obtained statistically in advance, this corrected matrix data is subjected to an inverse DCT operation to obtain image data, and this image data is converted to analog to obtain a video signal. , it is possible to correct the horizontal and vertical spatial frequencies of the video signal in a wide range (including high frequency components), and compared to the case of high frequency component correction for the horizontal spatial frequency, This has the effect of further improving the quality of the displayed image.

Furthermore, in the present invention, as in the above-mentioned correction of high-frequency components, noises obtained statistically in advance from the matrix data, such as noise patterns caused by tape hiss or satellite lines, are removed, so that the above-mentioned display image is corrected. Image quality can be further improved.・

[Brief explanation of the drawing]

FIG. 1 shows an embodiment of the present invention, and is a schematic block diagram of an image processing apparatus to which an image processing method is applied, and FIGS. 2 to 4 are diagrams for explaining the above image processing method. . In the figure, 1 is the A/I) conversion section, 2 is the DCT section, 3 is the tape hiss/noise pattern-number circuit, 4 is the first correction coefficient table section, 5 is the first calculation section, and 6 is the pattern matching circuit. ,
7 is the second correction coefficient section, 8 is the second calculation section, 9 is the inverse D
CT section, 10 is D//l: exchange section. Patent applicant: Fujitsu General Co., Ltd. Agent: Patent attorney: Takuya Ohara

Claims (1)

[Claims] (1) Digitally convert the video signal to image data,
The image data is subjected to a DCT operation, the matrix data of the DCT operation result is multiplied or subtracted by data obtained statistically in advance, and the high-frequency components are extracted using the correlation between the spatial frequencies in the horizontal and vertical directions. An image processing method comprising: correcting the corrected matrix data, performing an inverse DCT operation on the corrected matrix data to obtain image data, and converting the image data into an analog signal to obtain a video signal.