JPH03256489A - Luminance/chrominance signal separator circuit for television signal - Google Patents

Abstract

PURPOSE:To enable motion detection without an integrating operation in a time space area while improving detection accuracy by detecting motion information for the unit of a moving object area. CONSTITUTION:This luminance/chrominance signal separator circuit is composed of a moving object area motion detection circuit 1, motion coefficient setting circuit 2, still picture corresponding chrominance signal extraction circuit 3, moving image corresponding chrominance signal extraction circuit 4, coefficient loading circuit 5, adder circuit 6, BPF circuit 7, delay circuit 8 and subtraction circuit 9. When detection motion from a composite color television signal VS, an area including the motion of picture quality is extracted as the moving object area and the motion information is detected in respect to this moving object area. By using the obtained motion information, the motion is adaptively controlled. Thus, the integrating operation is not required in the time space area and the detection accuracy can be improved.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a motion-adaptive luminance signal and chrominance signal separation circuit for television signals.

[Conventional technology]

In YC separation processing, luminance signal component Y11 signal component C is separated from a composite color television signal.

In order to improve image quality by reducing cross-color and cross-luminance, motion-adaptive three-dimensional yc separation has come to be widely used.

In these motion-adaptive three-dimensional YC separations, components corresponding to still images obtained by inter-frame YC separation and components corresponding to moving images obtained by line-to-line YC separation are separated according to the movement of the image. By adaptively changing the mixing ratio, the luminance signal component Y22 and the signal component C are separated.

Image movement information necessary for this adaptive control is detected using a composite color television signal.

[Problem to be solved by the invention]

In the above-mentioned conventional technology, information on image movement is detected pixel by pixel based on the low-frequency component of the difference signal between one frame and the difference signal between two frames of a composite color television signal. For this reason, there is a problem in that motion detection is often omitted to determine that a moving image is a still image.

As a remedy for failure to detect motion, integration operations in the temporal direction using a spatio-temporal filter are sometimes performed, but in this case, when the video stops or dropouts occur, it is necessary to Since it takes time to switch to the still image mode, there is a problem in that unnatural image quality deterioration appears due to this.

An object of the present invention is to provide a circuit for separating high-quality luminance two-color signals by solving the problems of the prior art described above by detecting motion information with high motion detection accuracy and with extremely low detection omissions. Our goal is to provide the following.

[Means to solve the problem]

In order to achieve the above object, the present invention is a means for extracting a region including image quality motion as a moving body region in motion detection from a composite color television signal, and detecting motion information for this moving body region. The motion information obtained from the motion information is used to perform adaptive motion control, thereby achieving highly accurate luminance two-color signal separation processing.

[Effect]

In order to extract a moving part included in an image from a composite color television signal as a moving object region, and detect movement information for this moving object region, for example, When a frequency component (a frequency band on which no color signal is superimposed) or a difference signal between frames is detected in this moving object region, the entire moving object region is determined to be in motion. As a result, compared to pixel-by-pixel motion detection performed in the prior art, the probability of detection failure occurring can be made extremely low. Therefore, the integration operation in the time direction using a spatio-temporal filter as a remedy for detection failure can also be eliminated.

In other words, in the present invention, by detecting motion information in units of moving body regions, detection accuracy is high.
In addition, motion detection can be performed without integration operations in the spatiotemporal domain. As a result, the detection failure, which is a problem with conventional technology,
Alternatively, it becomes possible to perform motion-adaptive YC separation processing without deterioration caused by a remedy for detection failure, and realize separation of luminance signals and nine color signals with higher precision than in the prior art.

〔Example〕

Hereinafter, the present invention will be explained in detail with reference to Examples.

FIG. 1 shows the overall block configuration of an embodiment of a luminance nine-color signal separation circuit according to the present invention.

The composite color television signal Vs is sent to the moving body area motion detection circuit 1. Still image compatible color signal extraction circuit 3. Video compatible color signal extraction circuit 4. The signal is input to the delay circuit 8.

The moving body region motion detection circuit 1 extracts motion regions included in the image from the signal Vs as moving body regions, detects motion information in units of these regions, and outputs a motion information signal MI. .

Based on this motion information signal M1, the motion coefficient setting circuit 2 sets a motion coefficient k of the N level characteristic shown, for example, by the solid line or dotted line in FIG.

The still image corresponding color signal extraction circuit 3 extracts a color signal component CM suitable for a still image, for example, by performing an arithmetic operation between frames.

Further, the moving image compatible color signal extraction circuit 4 extracts color signal components CM suitable for moving images, for example, by arithmetic operations between lines.

The coefficient loading circuit 5 sets 1-k to the signal CM based on the motion coefficient obtained from the motion coefficient setting circuit 2.

Each signal CM is loaded with a coefficient k. and.

An adder circuit 6 adds the two signals, a BPF circuit 7 extracts a signal component in a predetermined color signal band, and generates a separated color signal component C.

On the other hand, from the composite color television signal whose delay has been adjusted by the delay circuit 8, the color signal component C is subtracted by the subtraction circuit 9 to generate a separated luminance signal component Y.

Note that the specific configuration of each block in this embodiment will be described in detail later.

Next, FIG. 2 shows the overall block configuration of another embodiment according to the present invention. This embodiment is characterized in that color signal components are extracted more precisely.

That is, in addition to still images and moving images, a circuit for extracting color signal components for semi-still images that move relatively slowly can be provided to perform separation processing that is more consistent with the characteristics of the movement.

From the composite color frequency vision signal VS, the quasi-still image compatible color signal extraction circuit 10 extracts a color signal component CSM suitable for slow movement by, for example, arithmetic operations between fields.
Extract.

These extracted color signal components are processed by a coefficient loading circuit 5 into motion coefficients having characteristics as shown in FIG. 4, for example.

k2. k, is subjected to coefficient weighting, and the separated color signal component C is generated by adding the coefficients. Then, the color signal component C is subtracted from the composite color television signal to generate a separated luminance signal component Y.

Next, FIG. 5 shows the overall block configuration of an embodiment in which the present invention is applied to a decoded color television signal Vsc that is compatible with the current system described in, for example, Japanese Patent Laid-Open No. 59-171387. In this case, separation processing of the superimposed high-definition signal is further added.

That is, the high-precision signal extraction circuit 11 extracts the first . The high-definition signal component superimposed on the third quadrant is extracted. Note that this high-definition signal is not superimposed on a moving image, so the motion coefficient kH of the characteristic shown by the dotted line in FIG.
Components in a predetermined high-definition signal band are extracted to separate high-definition signal components SH.

On the other hand, regarding the color signal, in a still image, the color signal component is the second digit in the time-vertical frequency domain. Since it is defined in the fourth quadrant, the still image corresponding color signal extraction circuit 13 extracts the color signal component C3)I corresponding to this. Then, C8H and CM are mixed and added according to the motion coefficient k to generate a separated color signal component C.

and a compatible composite color television signal V
A separated luminance signal Y is generated by subtracting the signal components C and SH from sc.

This completes the explanation of the overall block configuration, and next, the constituent elements of each block will be described in detail based on embodiments.

An embodiment of the moving body area motion detection circuit 1 is shown in FIG.

In this embodiment, the luminance component extraction circuit 14
．． Object region extraction circuit 15. Frame difference signal extraction circuit 1
6. This is realized by combining the motion information extraction circuit 17.

The luminance component extraction circuit 14 has a function of extracting a luminance component signal Y necessary for extracting an object region from the composite color television signal Vs (Vsc), and has an embodiment shown in FIGS. 8 to 10. This can be achieved with

In the embodiment shown in FIG. 8, a one-pixel delay circuit 18 provides a time delay corresponding to inversion of the phase of the color subcarrier of the composite color television signal, and an adder circuit 6 adds these two signals to produce a luminance component signal Y. Extract the components of

In the embodiment shown in FIG. 9, a signal is delayed by a time corresponding to one line period in a 51-line delay circuit 19 by utilizing the property that the phase of the color subcarrier of a composite color television signal is inverted for each line. By adding these, the component of the luminance component signal Y is extracted.

The embodiment shown in FIG. 10 utilizes the property that the phase of the color subcarrier of the composite color television signal is inverted for each frame, and uses the characteristic that the phase of the color subcarrier of the composite color television signal is inverted for each frame to generate a signal delayed by a time corresponding to one frame in the one frame delay circuit 20. The components of the luminance component signal Y are extracted by addition.

The object region extraction circuit 15 has a function of extracting an image pattern from the luminance component signal Y and extracting the image pattern portion as one object region based on the moss.
This can be realized with the configuration example shown in the figure.

Luminance component signal Y and 1 pixel delay circuit 2
1 and the signal delayed by one pixel is subtracted by a subtraction circuit 9 to extract a signal ○H corresponding to a vertical pattern included in the image.

Similarly, a signal ○V corresponding to the horizontal pattern included in the image is extracted from the difference with the signal delayed by one line in the one line delay circuit 19.

These signals are subjected to binarization processing in a binarization circuit 22, where they are set to 1 if they exceed a threshold value ±ΔTh, and 0 otherwise.
Generate signal ○H2○V.

Then, the smoothing circuit 23 removes noise components.

In addition, smoothing processing is performed to extract the image pattern portion as one continuous region. This processing can be easily realized using a ROM or the like.

The smoothed signal is subjected to an OR operation in an OR circuit 24 to generate an object area signal BS. Note that this signal includes information on both stationary and moving object areas.

The frame difference signal extraction circuit 16 has a function of taking a difference signal between frames of a composite color television signal and extracting the motion of an image, and can be realized by the embodiment shown in FIGS. 12 to 14.

FIG. 12 utilizes the property that the phases of the color subcarriers of the composite color television signal match every two frames, and the composite color television signal and the two-frame delay circuit 25 correspond to two frames. The difference signal FD is extracted by subtracting the difference between the time-delayed signal in the subtraction circuit 9.

FIG. 13 utilizes the property that there is no color signal in the low frequency band of a composite color television signal, and passes the difference signal from the signal delayed by one frame equivalent in the one frame delay circuit 20 to the LPF circuit 26. .

The low frequency component is extracted as a difference signal FD.

In FIG. 14, a difference signal FD is extracted from the low frequency component of the difference signal between two frames and the difference signal between one frame of a composite color television signal. That is, L
One-frame difference signal F D obtained from the PF circuit 26
1F and the two-frame difference signal FD2F, a maximum absolute value selection circuit 27 selects the signal with the larger absolute value of the signal component and extracts it as the difference signal FD.

In other words, FDiFl-IFD, Fl is FD8FTI.
When FDiFl<1FD2F+, F_D2F is extracted as the signal FD.

The motion information extraction circuit 17 receives an object area signal BS.

It has a function of generating a motion information signal MI based on the difference signal FD, and can be realized by the embodiment shown in FIG. 15.

One of the differential signals FD is as shown in FIG.
A binarization circuit 22 binarizes the signal, and a smoothing circuit 23 performs smoothing processing to generate a signal FD2. This signal and the object area signal ○BS are input to the moving body area determination circuit 28,
An object region containing movement is extracted as a moving object region signal MOBS from the correlation between both signals.

The principle of this operation is shown in FIG. In the object region (■) of the signal OBS, since no difference is produced in the signal FD2 during periods a and b, this period is determined to be a stationary area such as a background. On the other hand, it is determined that there is movement during a period in which a difference is detected in the signal FD2, and the period corresponding to this is regarded as a moving body region, and the signal MOBS is set to 1, for example.

Since there is no signal FD2 in the @ object area of signal OBS,
This object region is determined to be stationary.

Furthermore, in the object region of the signal OBS, a difference is also detected in the signal FD2 in agreement with this.

This entire area is determined to be a moving body area.

These operational functions can be realized by a combination of logic circuits.

Next, the other one of the difference signals FD is subjected to absolute value quantization by 1, for example, with the characteristics shown in FIG. 17 in the absolute value quantization circuit 29 to generate a signal QFD.

The motion information generation circuit 30 generates moving body area signals MOBS,
It has a function of generating a motion information signal MI based on the signal QFD and the signal QFD, and is realized by an embodiment having the configuration shown in FIG.

The switch circuit 32 is controlled by the signal MOBS, which
When this is 1 (when it is not a moving body region), the input β is selected, and when it is 1, that is, when it is a moving body region, the input α is selected and output.

Further, the MAX selection circuit 31 selects the highest value among the input signals and outputs the selected signal.

Further, the time axis inversion circuit 34 performs a time axis inversion process on a line-by-line basis, for example, and outputs the output signal QFD'
, MOBS' is a time series signal in which the last data of one line is the beginning and the first data of one line is the last.

Next, this operation will be explained. In areas other than the moving body area, the switch circuit 32 is connected to the input β, so the signal Q
FD outputs signal MI as it is.

The system becomes MI2 and detects motion information in the same way as in the prior art.

On the other hand, in the moving body region, the switch circuit 32 inputs α
, and forms a feedback loop with the MAY selection circuit 31. And in the moving body region, the signal QFD
The maximum value of is generated as the output signal M 11°M■. As a result, in the part of the moving body region where the signal QFD is zero, the value that was the maximum before that part is held.

In other words, as in the present invention, in the moving body region, the signal Q
By using the maximum value of FD as movement information,
It becomes possible to significantly suppress the occurrence of detection failures, which is a problem with conventional techniques.

In addition, in order to improve the detection accuracy of motion information, the signal MI2 obtained from the signal sequence with the time axis inverted is again converted into the original time series signal by inverting the time axis, and the delay is adjusted by the delay circuit 33. The maximum value of both the MI□ signal and the MI□ signal is finally extracted as the motion information signal MI. Therefore, in some cases, signal processing based on time axis inversion may be omitted.

This concludes the description of the moving body region motion detection circuit 1. Next, an embodiment of the configuration of the motion coefficient setting circuit 2 shown in FIGS. 19 to 21 will be described.

FIG. 19 shows this block configuration. Motion information signal MI
The smoothing circuit 35 performs smoothing processing in the horizontal, vertical, or both areas to obtain the smoothed motion signal MIAV.
generate. Then, the motion coefficient generation circuit 36 generates a signal M
Depending on the level of IAV, for example, the motion coefficient k (k□, k2.
k3. generates kH). These can be easily realized using a ROM or the like.

FIG. 20 shows a configuration example of a smoothing circuit that performs smoothing processing in the horizontal direction. For the signal sequence delayed by the one-pixel delay circuit 21, the coefficient a. , al.

..., aN are coefficient-loaded, and these are added to generate a smoothed motion signal MIAV. Coefficient a. ,..., a
N may be set so as to obtain desired smoothing characteristics.

FIG. 21 shows an example of a configuration in which smoothing processing is performed in two-dimensional horizontal and vertical regions. For the signal series delayed by the 1-line delay circuit 19, the coefficients avo l..., a
Smoothing in the vertical direction is performed by adding a load to vm. Furthermore, a coefficient a is applied to the signal sequence delayed by the one-pixel delay circuit 21. , . . . , aN is added as a load to perform horizontal smoothing processing.

This concludes the explanation of the motion coefficient setting circuit 2, and next,
An embodiment of the color signal extraction circuit will be described with reference to FIGS. 22 to 26.

FIG. 22 shows an example of the configuration of the still image compatible color signal extraction circuit 3. For the signal sequence that is frame-delayed by the one-frame delay circuit 20, coefficients -1/4.1/2. -1/4 is weighted and the color signal component CS is extracted using a so-called frame comb filter.

FIG. 23 shows an example of the configuration of a still image compatible color signal extraction circuit 13 suitable for a composite color television signal (described in Japanese Patent Laid-Open No. 171.387/1987) that is compatible with the current system. .

The color signal component CM obtained from the frame comb filter is
The signal delayed by a time equivalent to 262 lines by the 262 line delay circuit 37 is added to obtain the second
．． Extract the color signal component CMH in the fourth quadrant.

FIG. 24 shows an example of the configuration of the moving image compatible color signal extraction circuit 4. 1 line delay circuit 19. A line comb type filter is constructed by combining the coefficient loading circuit 5.

To extract a color signal component CM suitable for a moving image.

FIG. 25 shows another example of the configuration of the moving image compatible color signal extraction circuit 4. In FIG.

Color signal component CLI extracted by line comb filter
N and the color signal component CBPF extracted in a horizontal one-dimensional frequency band by the BPF circuit 39 are selectively outputted by a selection circuit 40 according to the horizontal pattern of the image. That is, the subtraction circuit 9. According to the horizontal pattern information of the image obtained by the LPF circuit 38, the CB
PF, otherwise the CLIN signal is the 8-power signal CM.

FIG. 26 shows an example of the configuration of the color signal extraction circuit 10 for the quasi-stationary side. 262 line delay circuit 37° 263 line delay circuit 41. A combination of coefficient loading circuits 5 constitutes a field comb filter, and color signal components C suitable for quasi-still images are generated.
Extract SM.

In addition to the YC separation processing described above, the motion detection according to the present invention can also be applied to motion adaptive processing such as scanning line interpolation. An example of this is shown in FIG.

Luminance signal Y separated from composite color television signal
An interpolated scanning line generation circuit 43 for still images and an interpolated scanning line generation circuit 44 for moving images generate interpolated scanning line signals YS+YM suitable for still images and moving images, respectively.

On the other hand, the motion information detection circuit 42 extracts motion information from the composite color television signal in the manner described above, and generates a motion coefficient k. This can also be used in common with the YC separation circuit described above.

The coefficient loading circuit 5 loads YS with a coefficient of 1- and YM with a coefficient of k in accordance with this motion coefficient k, and adds the two in an adder circuit 6 to generate an interpolated scanning line signal Yrp.

The luminance signal and the interpolated scanning signal V whose delay has been adjusted by the delay circuit 45
The IP is converted to 172 on the time axis by the time axis conversion circuit 46.
Compression and time-series rearrangement and multiplexing are performed to obtain a signal in a desired sequential scanning format.

〔Effect of the invention〕

According to the present invention, motion detection failures are extremely rare, and there is no need for remedies for detection failures such as integration operations in the spatiotemporal domain.
Motion adaptive YC using motion information signals with high detection accuracy
Since separate signal processing can be realized, it becomes possible to optimally separate the luminance signal and color signal that match the motion, which has a great effect on improving the quality of television images.

It is clear that the present invention is effective not only for existing composite color television signals but also for composite color television signals that are compatible with the present composite color television signals.

It is also clear that the motion information obtained by the present invention can also be used in motion adaptive processing rings such as scan line interpolation.

Further, in this embodiment, although the 1-frame delay circuit and the like are shown in separate configurations for each function, it is clear that these can be used in common.

Figures 5 and 6 show the overall block configuration and characteristic diagrams of an embodiment in which the present invention is applied to a compatible composite color television signal.

FIGS. 7 to 18 show the overall block configuration of the moving body area motion detection circuit used in the present invention, an example of each block, and
and - characteristic diagram.

Further, FIGS. 19 to 21 show an example of a motion coefficient setting circuit, and FIGS. 22 to 23 show a still image compatible color signal extraction circuit,
Figures 24 to 25 are video-compatible color signal extraction circuits;
The figure shows an embodiment of a color signal extraction circuit for the quasi-stationary side.

1. Moving body region motion detection circuit, 2. Motion coefficient setting circuit, 3. Still image compatible color signal extraction circuit, 4. Video compatible color signal extraction circuit, 5. Coefficient loading circuit, 6. Addition circuit, 7...BPF circuit, 8...Delay circuit, 9...Subtraction circuit, 10...Quasi-stationary side compatible color signal extraction circuit, 11
...High-definition signal extraction circuit, 12...Filter circuit, 13
...Still image compatible color signal extraction circuit, 14.. Luminance component extraction circuit, 15.. Object region extraction circuit, 16.
... Frame difference signal extraction circuit, 17... Motion information extraction circuit, 18... Pixel delay circuit, 19... 1 line delay circuit, 20... l frame delay circuit, 21... 1
Pixel delay circuit, 22... Binarization circuit, 23... Smoothing circuit, 2 = OR circuit, 25. 2 frame delay circuit, 26... LPF circuit, 27... Absolute value maximum selection Circuit, 28 Moving body region determination circuit, 29 N@electronic circuit 30...Movement information generation circuit, 31...MAX
Selection circuit, 32... Switch circuit, 33... Delay circuit, 34... Time axis inversion circuit, 35... Smoothing circuit,
36... Motion coefficient generation circuit, 37... 262 line delay circuit, 38... LPF circuit, 39... BPF circuit, 40... Selection circuit, 41... 263 line delay circuit,
42... Motion information detection circuit, 43... Interpolated scanning line generation circuit for still images, 44... Interpolated scanning line generation circuit for moving images, 45... Delay circuit, 46... Time axis conversion time 11 Fig. 2 Fig. Nest/6 Fig. % Noah rhyme zl Fig. 7 Z in 2 42# times y valley, ■ 7 Fig.

Claims (1)

[Scope of Claims] 1. In a luminance/chrominance signal separation circuit that separates a luminance signal Y and a color signal C from a composite color television signal, means for extracting an image pattern region from the composite color television signal as an object region; A means for extracting a difference signal between different frames of a composite color television signal, a means for separating an object region containing movement from the object region and the difference signal as a moving object region and extracting information on the movement, and the above extraction. It has means for generating a motion coefficient k from the information obtained, means for extracting a color signal component C_S suitable for a still image, and means for extracting a color signal component C_M suitable for a moving image. ) A brightness/chrominance signal separation circuit for a television signal, characterized in that it separates C_S+kC_M as a color signal C. 2. The brightness of the television signal according to claim 1, wherein the composite color television signal is a composite color television signal that is compatible with current television.
Color signal separation circuit.