JPH0738726B2 - Luminance color signal separation circuit - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion-adaptive luminance color signal separation circuit for television signals, and particularly to a luminance suitable for a high-definition television system compatible with the current television system. The present invention relates to a color signal separation circuit.

[Background of the Invention]

The signal band and the superposition of color signals are the same as those of the current television signal system, and it is possible to send and receive high resolution images.
As a so-called compatible high definition television system, a system described in Japanese Patent Application No. 58-44238 is known.

In this system, as shown in FIG. 1, the color signal C is arranged in the second and fourth quadrants on the vertical two-dimensional frequency in the horizontal frequency region of 2 MHz or more as in the current television system. There is. Further, the high-definition signal Y _H ′ is multiplexed and transmitted in the first and third quadrants on the vertical two-dimensional frequency in the horizontal frequency band where the color signal C is superimposed.

In this method, it is necessary to separate the luminance signal Y, the color signal C, and the high definition signal Y _H ′ on the receiving side. In this separation, so-called motion adaptation processing, in which the separation characteristic is changed according to the motion, is performed, so that separation with less crosstalk can be realized. In this method, since the transmitting side also performs signal processing, the transmitting side detects motion information and multiplexes this as mode information to the image signal for transmission, and the receiving side uses this mode information for motion adaptation. Methods for determining processing parameters are known. However, in this method, since it is impossible to give the mode information to all the pixels due to the limitation of the transmission band, the control is performed in a rough range such as a field unit or a block unit. .

For this reason, even if a part of the block is moving, the entire block is processed as a motion mode.
There is a problem that the separation characteristics deteriorate.

[Object of the Invention]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a luminance / color signal separation circuit for motion adaptation processing which detects a motion from a transmitted high definition television signal and controls a separation parameter.

[Outline of Invention]

Various methods have been devised for detecting motion from a received signal with respect to a current television signal, but in a compatible high definition television signal, in addition to the luminance signal Y and the color signal C, higher The fine signal Y _H ′ is newly superimposed. Therefore, in the conventional motion detection, there is a possibility that the high definition signal Y _H ′ is erroneously detected as a motion signal.

In the present invention, as shown in FIG. 2, a new motion detection is performed in which a shaded area is determined as a still image area in a time and vertical two-dimensional frequency area. The reason for this is that in a compatible high definition television signal, the luminance signal Y has a signal spectrum component on the ν axis and the color signal C and the high definition signal Y _H ′ have a signal spectrum component on a time frequency = ± 15 Hz for a still image. Because it exists. Therefore, the signal spectrum of the high-definition television signal is detected in the time and vertical frequency regions, and if it exists only in the shaded region, a still image,
If the image is also present in a region other than the diagonal line, it is possible to detect the motion for each pixel by determining that the image is a moving image.

Example of Invention

 An embodiment of the present invention will be described below with reference to FIG.

The high definition television signal has 26 signals as shown in FIG.
A delay circuit in which a two-line delay circuit 1, a one-line delay circuit 2, and a 261-line delay circuit 3 are combined makes
As shown in, X ₀ , X _-1 which is one line away from this X ₀
And X ₁ , 262 lines apart X _-262 and X ₂₆₂ , 263 lines apart X
_-263 and X _263, and constitutes a 525-line signal of the X _-525 and X ₅₂₅ away. These signals are supplied to the motion detection circuit 4 and the separation circuit 5. The motion detection circuit 4 detects motion from these signals and determines the separation parameter characteristic. In the separation circuit 5, according to this parameter characteristic,
The luminance signal Y, the color signal C, and the high definition signal Y _H ′ are separated.

Next, a configuration example of the motion detection circuit 4 will be described with reference to FIG.

Motion detection is performed by a plurality of types of filters. The spatio-temporal filter 6 carries out a calculation of X ₅₂₅ -X _-525 and takes a difference between so-called two frames. As shown in FIG. 3, in X ₅₂₅ and X _-525 , the _phases of the color signal C and the high-definition signal Y _H ′ are the same, and this characteristic is that the shaded area and hatch shown in FIG. 5 (a). The area of the part is used as the pass band.

On the other hand, the space-time filter 9 is one that takes a low frequency component of X ₀ -X _-525 with a horizontal frequency μ of, for example, 2 MHz or less. That is, it is a low-frequency component with a difference between 1 frames, and the characteristic is that the shaded area and hatched area as shown in FIG.

Moreover, the space-time filter 11 is (X ₂₆₂ −X ₋₂₆₃ ) + (X ₂₆₃
The calculation of −X ₋₂₆₂ ) is performed to obtain the one-line sum of the one-frame difference, and the characteristic is that the shaded area and the hatched area in FIG. 5C are the passbands.

Furthermore, the spatiotemporal filter 13 is (X ₀ −X ₋₂₆₂ ) + (X ₅₂₅
The characteristic shown in FIG. 5D is realized by the calculation of −X ₂₆₃ ).

Since a television signal forms a signal by scanning, it corresponds to a kind of sampling in the time and vertical frequency domain. In the time and vertical frequency domain, the time frequency f =
30Hz, vertical frequency A sampling frequency f _IS is generated at the point, and a signal spectrum is also generated around this point. Therefore, FIG. 5 (a)
In (d) to (d), the signal in the shaded portion, which is the original signal, is generated as a folding component in the hatch portion at the sampling frequency f _IS .

The output signals of the spatiotemporal filters are set to 0 in the threshold circuits 7, 10, 12, and 14 when the absolute value of the filter output signal is less than a certain threshold Δth, and set to 1 when the absolute value exceeds Δth. It occurs. These threshold circuits are RO
It can be easily realized with M etc.

Now, of the output signals of the threshold circuit, the space-time filter 11,
MD4 and MD5 corresponding to the spatio-temporal filter 13 may output 1 even in a still image, as is apparent from FIGS. 5 (c) and 5 (d). For example, the output of MD4 may be 1 for an image in which the color changes in the vertical direction, and MD5 may be 1 for an image with a high vertical component of the luminance signal even for a still image.

Therefore, in the motion / motion identification circuit 8, the space-time filter 6
Based on the signal MD2 of the difference between the two frames, a moving image or a still image is first identified in a large area. This discrimination is performed, for example, as shown in FIG. 6, by providing a judgment area composed of a plurality of peripheral pixels including the pixel of interest, and using the MD2 value of the pixels in this area. For example, if all the MD2 of the pixels in the determination area are 0, the pixel of interest is identified as still and the output MD1 is 0.
To occur. On the other hand, if the MD2 of the pixels in the determination area is 1, the pixel of interest is identified as a moving image, and MD1 is 1. In addition, as shown in FIG. 5 (a), f = ± 15 Hz,
This is because motion near ± 30 Hz may be missed, so detection of motion can be reduced by detecting motion in a large area.

The motion determination circuit 15 determines the motion mode using the information of MD1, ... MD5 obtained by the above procedure. In this embodiment, the still mode is used in the case of stillness, the quasi-stationary mode is used in the case of relatively slow movement having a low time frequency f, and the moving image mode is used in the case of a rapidly changing movement having a high time frequency f. Making a decision. As shown in FIG. 4 (b), for example, when MD1 is 0, this pixel determines the stationary mode. As a result, as described above, it is possible to correctly determine the still mode even in the area where MD4 and MD5 are 1 when a still image.

On the other hand, when MD1 is 1, the mode is determined by the information of MD2 to MD5. That is, when MD2 to MD5 are all 0,
Since there is no signal spectrum in each of the passbands shown by the hatched portion and the hatched portion in FIGS. 5A to 5D, the passbands in FIGS. 5A to 5D are overlapped. If you look at it,
When the horizontal frequency μ is less than 2MHz, the time frequency f is 15/4
Vertical frequency ν below Hz Only in the following areas, and when the horizontal frequency μ is 2MHz or more, the above area and the time frequency f are 45 / 4Hz to 74 / 4Hz and the vertical frequency ν is This corresponds to the case where the signal spectrum exists only in the area of (1) and the signal spectrum exists only in the area of the diagonal line shown in FIG. 2, and this pixel is determined to be the stationary mode. When MD2 is 1 and the rest are all 0, the horizontal frequency μ is determined by superimposing the passbands of FIGS. 5A to 5D and applying the presence or absence of the signal spectrum in each passband. Below 2MHz, time frequency f is 15 / 4Hz to 15 / 2Hz and vertical frequency ν
But In the following region, and when the horizontal frequency μ is 2MHz or more, the above region and the time frequency f are 15 / 2Hz to 45 / 4Hz and the vertical frequency ν is Since the signal spectrum exists in the region of Fig. 2, this corresponds to the case where the signal spectrum also exists in the region other than the diagonal line shown in Fig. 2, and in the image with the low vertical frequency ν, the time frequency f is 15 / 4Hz to 15 / 4Hz. This pixel is determined to be in the quasi-stationary mode because it corresponds to the case of detecting the movement of / Hz.

Also, if any of MD3, MD4, or MD5 is 1, from the same viewpoint as above, it corresponds to the case where the signal spectrum exists also in the area other than the hatched area shown in FIG. Frequency f is 15 / 2Hz ~ 45 / 2Hz, 15 / 2Hz ~
Since this corresponds to the case where a motion of 45 / 2Hz or 45 / 2Hz to 105 / 4Hz is detected, this pixel is determined to be the moving image mode.

Next, FIG. 7 shows a configuration example of the separation circuit 5 shown in FIG.

The frame separation circuit 16 The C & Y _H ′ component is extracted by the calculation of This signal is C
And Y _H ′ components are extracted. This signal, for separating C and Y _H 'components are added to the 262H delay circuit 1. Subtraction circuit 20
Then, by subtracting the output signal of the 262H delay circuit,
As shown in FIG. 1, the Y _H ′ signals in the first and third quadrants in the time and vertical frequency domain are separated. On the other hand, in the adder circuit 21, 262H
By adding the output signals of the delay circuit, C in the second and fourth quadrants
Separate the signals. This becomes the C separation signal between frames.

In the inter-field separation circuit 17, Is calculated. Then, addition with the signal of the 262H delay circuit 1 is performed by the adder circuit 19 to separate the C signals in the second and fourth quadrants. This becomes the C separation signal between the fields.

The line separation circuit 18 is The C separation signal between lines is obtained by the calculation of.

The C signal selection circuit 22 selects the C signal according to each separation mode, as shown in FIG. 7 (b). Then, through the horizontal band pass filter 23, a C signal having a horizontal frequency μ of 2 MHz or more is extracted. Note that the selection circuit 24, in the stationary mode and when MD1 is 0, determines that the signal is completely stationary and selects a signal that does not pass through the horizontal band pass filter. On the other hand, MD
When 1 is 1, the signal passed through the horizontal bandpass filter is the C signal.

Further, the Y _H ′ selection circuit 25 outputs the signal of the subtraction circuit 20 as a Y _H ′ signal when MD1 is 0 and in the stationary mode.
On the other hand, when MD1 is 1, the Y _H ′ signal outputs 0.

Finally, the subtraction circuit 26 subtracts the Y _H ′ signal and the C signal from the X ₀ signal to separate the luminance signal Y.

〔The invention's effect〕

According to the present invention, since it is possible to detect motion with less detection error, Y, C, Y _H ′ separation of compatible high-definition television signals can be separated pixel by pixel. Is extremely effective for

In the embodiment, the determination of the separation mode, the separation characteristic, and the like have been described in the case where the weighting coefficient has a binary value of 0 or 1, but three or more weighting coefficients are prepared and the parameters are continuously switched. It is also possible.

Further, in the embodiment, the explanation has been limited to the high-definition television signal, but it goes without saying that it is also effective for the current NTSC signal.

Further, the determination result of the separation mode obtained in the embodiment can be used for the motion adaptation processing of the scanning line interpolation when the scanning frequency of the current television is doubled and the progressive scanning is displayed.

It should be noted that the determination of the separation mode, the separation characteristics, and the like can be weighted including changes in the time direction.

[Brief description of drawings]

FIG. 1 is a compatible high-definition television time, vertical frequency domain signal spectrum, FIG. 2 is an explanatory view of the principle of motion detection according to the present invention, and FIG. 3 is a block diagram of an apparatus according to an embodiment of the present invention. And a signal configuration diagram, FIG. 4 is a block diagram and a signal state diagram showing the configuration of the motion detection circuit in the embodiment of the present invention, FIG. 5 is a characteristic diagram of each signal, FIG. 6 is a schematic diagram of a determination region, FIG. 7 is a block diagram and a separation characteristic diagram showing the configuration of the separation circuit. 1 …… 262H delay circuit, 2 …… 1H delay circuit, 3 …… 261H delay circuit, 4 …… Motion detection circuit, 5 …… Separation circuit, 6,9,1
1,13 …… Spatio-temporal filter, 7,10,12,14 …… Threshold circuit, 8
...... Motion identification circuit, 15 ...... Motion judgment circuit, 16 ...... Frame separation circuit, 17 ...... Field separation circuit, 18 ...... Line separation circuit, 19, 21 ...... Addition circuit, 20, 26 ...... subtraction circuit, 22 ...... C signal selection circuit, 23 ...... horizontal bandpass filter, 24 ...... selection circuit, 25 ...... Y _H 'selection circuit.

Claims (2)

[Claims] 1. A composite color television signal as an input,
A delay unit including a plurality of delay circuits, a motion detection unit that receives the composite color television signal and a plurality of delay signals output from the delay unit, detects a motion, and outputs a separation parameter, and the composite color A television signal, a plurality of delay signals output from the delay unit and the separation parameter as an input, and a separation unit that adaptively changes the separation characteristics according to the separation parameter, the motion detection unit, A first spatio-temporal filter which is a stop band where the time frequency in the vertical and two-dimensional frequency domain f = 0 Hz and the point on f = ± 15 Hz are zero, and the composite color television signal includes a color signal In the horizontal frequency band in which the components are superimposed, the time, the vertical two-dimensional frequency domain time frequency and vertical frequency Of the two spatio-temporal filter groups, each of which is a stop band whose zero point is zero and which has different pass bands from each other, and each of the first spatio-temporal filter and the second spatio-temporal filter group. A motion adaptive luminance / color signal separation circuit comprising: a motion determination unit that adaptively changes and outputs a separation parameter according to a combination of output results. 2. The composite color television signal is a color signal in the second and fourth quadrants of the time and vertical two-dimensional frequency domain,
The motion adaptive luminance / color signal separation circuit according to claim 1, characterized in that it has a signal spectrum of a high definition signal in the first and third quadrants.