JPH04157893A - Motion detecting circuit for pal color television signal - Google Patents

Abstract

PURPOSE:To prevent the deterioration of a picture quality by sampling a PAL color television signal by a sampling frequency signal which is four folds of a chrominance subcarrier frequency, and also, matches with a phase of a color burst signal, and obtaining a motion signal by synthesizing a first and a second motion signals. CONSTITUTION:A clock signal for sampling is generated by a clock generating circuit 3, and set to a phase which is four folds of a frequency fSC of a chrominance subcarrier, and also, matches with the phase of a color burst signal. Arithmetic circuits 8 and 9 generate a difference signal CL between two frames and a difference signal CF between two frames, respectively. Subsequently, the CL signal and the CF signal are synthesized in a mixing circuit 10, and a chrominance signal (C signal) of a high quality processed correctly in accordance with a motion of an image is obtained. Since a PAL color television (P) signal is obtained by synthesizing a Y signal and a C signal, when the sampled P signal is subjected to delay adjustment by a delay circuit 12, and the C signal is subtracted by an arithmetic circuit 13, the Y signal of a high quality processed adaptively against the motion is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a motion detection circuit for separating a luminance signal and a carrier color signal from a PAL composite color television signal according to the motion of nine images.

[Conventional technology]

P A L (Phase Alternation
by Lin5) system color television signal (hereinafter abbreviated as P signal) is composed of a composite signal in which a carrier color signal (hereinafter abbreviated as C signal) is superimposed on a luminance signal (hereinafter abbreviated as Y signal). The C signal superimposed on the signal is obtained by orthogonally modulating two types of color difference signals (U signal and V signal) using a color printing carrier wave.
It is frequency multiplexed within the Y signal station area.

Red and green are used to play the P signal on a receiver.

It must be converted into a video signal of the three primary colors of blue.

As pre-processing to obtain video signals of the three primary colors, separation of the Y signal and C signal (hereinafter referred to as RYC separation) is essential.

According to the standard, the following relational expression holds true for the P signal.

P=Yu−sin (2πfsc” t)±v Red c
os (2πfscet) ・Curved surface (1) f wings =
(284-1/4+1/625)・fH・・・・・・(
2) fH=625・fF・・・・・・・・・・・・
...Song...Song...Listening... (3) Here, +fsc is the frequency of the color subcarrier that modulates the U and V signals, t is time, fs is the horizontal scanning frequency, and fF is the frame frequency. represent. The second and third terms on the right side of equation (1) represent the C signal, and the same magnitude signifier indicates that the polarity of the V signal is reversed line by line. In addition, equation (2) shows that the phase of fsc is offset by π/2 [radl with respect to the line period, and further substituting equation (3) into equation (2), fsc is It can be seen that it is also an offset of π/2 (radl) with respect to the frame period. Therefore, when looking at the line unit in the field, the phase of the C signal is reversed every 2 line periods and is the same every 4 line periods. .Also, in terms of frames, the polarity is reversed every two frames.

The relationship is of the same polarity at a cycle of 4 frames.

By utilizing the phase relationship of the C signal, the difference signal between two frames of the P signal becomes a complete C signal in a portion where the video is stationary, so that it can be completely separated from the corresponding Y signal. However, when the video is moving, the difference signal between two frames of the P signal does not become the C signal, so the corresponding Y
It cannot be separated from the signal. Therefore, where the video is moving, the difference signal between two lines in the same field can be used as the C signal and separated from the corresponding Y signal so that crosstalk between the Y signal and C signal is minimized. . Therefore, efficient YC separation can be realized by detecting the movement of the image from the P signal and changing the signal processing form of YC separation between two frames and between two lines according to the movement.

Such YC separation technology is described in 2 literatures I.E.
Transactions on Consumer Electronics, IEEE 32, 3 (August 1986), pp. 241-250 (IEEE Trans.
Consumer Electronics.

GE-32,3 (August 1986), pp.
241-250) [hereinafter referred to as the first reference document]. This conventional technology extracts the C signal from the difference between two frames of the P signal in a still part of the video.
In the moving part of the image, the C signal is extracted from the difference between the two lines of the P signal, and the mixing ratio of both C signals is adjusted according to the movement of the two images, changing the signal processing form of YC separation. .

FIG. 4 shows a configuration example of a YC separation circuit realizing the above-mentioned conventional technique. The P signal applied to the input terminal 1 passes through the first line memory circuit 4 and the second line memory circuit 5,
It is subject to a delay of 2 HD periods (HD represents 1 horizontal scanning period). Then, in the calculation circuit 8, subtraction is performed with the P signal which is not subjected to delay, and the difference signal, that is, the carrier color signal C is outputted. On the other hand, the P signal applied to the input terminal 1 at the same time passes through the first frame memory circuit 6 and the second frame memory circuit 7 and is delayed by 2 FD periods (FD represents one frame period). and.

The arithmetic circuit 9 performs subtraction with the P signal which is not subject to delay, and outputs a carrier color signal CF. The two carrier color signals C and CF are mixed in a mixing circuit 10 to form a new carrier color signal C. The mix ratio is determined by a motion control coefficient k generated by the motion detection circuit 20, depending on the state of image motion. That is.

C=-C+, + (1-k)・CF However, 0≦≦1. The C signal produced by the mixing circuit 10 is guided to an output terminal 15 via a two-color bandpass filter 11. Also, a part of the C signal that has passed through the nine-color bandpass filter 11 is applied to the arithmetic circuit 13, and subtracted from the P signal whose delay has been adjusted in the delay circuit 12 to generate the Y signal.

It leads to the output terminal 17. Naturally, the mixing circuit 10
Since the C signal formed in is adapted to the movement of the image, the Y signal is also adapted to the movement.

Now, in FIG. 4, the motion control coefficients generated by the motion detection circuit 20 affect the YC separation characteristics. If the motion detection is incorrect, the YC separation will fail, resulting in deterioration of the two-image quality. The first reference document describes, as a motion detection method, a difference signal between the current frame and the frame immediately before it (hereinafter referred to as a difference signal between one frame). ),
Alternatively, a difference signal (hereinafter referred to as a difference signal between 4 frames) between the current frame and the frame 4 frames before it (that is, 3 frames in between) may be used. It is shown.

FIG. 5 shows the configuration of the motion detection circuit 20. The P signal applied to input terminal 1 is delayed by one frame period or four frame periods by delay circuit 50. Then, the arithmetic circuit 51 performs subtraction with the P signal that is not subject to delay, thereby extracting a motion signal. next,
The obtained motion signal is rectified in a rectifier 52 and converted into a unipolar signal. Furthermore, the spatial filter circuit 5
3, only the low frequency components of the rectified motion signal are extracted. After that, in addition to the nonlinear conversion circuit 54.

A motion control coefficient k is generated by amplifying the signal and performing nonlinear processing.

On the other hand, regarding motion detection of the P signal, if the P signal is sampled at four times the color subcarrier frequency (4fac) and at a position where the phase matches the phase of the color burst signal, one frame of the P signal JP-A-2-108 shows that by creating a difference between pixels, it is possible to extract motion signals for every other pixel.
390 (hereinafter)

2nd reference].

FIG. 6(a) shows the configuration of the disclosed motion detection circuit. The P signal applied to input terminal 1 is A/
The signal is sampled by a sampling circuit 2 having a D (analog-to-digital) conversion function.

Sampling at this time is performed using a predetermined clock generated by the clock generation circuit 3, that is, the frequency fsc of the color subcarrier.
This is done using a signal that matches the phase of the color burst signal. When the sampled P signal is delayed by one frame period in the IFD memory circuit 60 and subtracted from the P signal without delay by the arithmetic circuit 61, the output signal S is as shown in FIG. As shown in b), the carrier color signal C and the motion signal m appear alternately in units of one clock. Therefore, the output signal S is transmitted to the distribution circuit 62 by [1/2 fsc
), the motion signal m can be separated and extracted by sampling with a control signal of period. The separated and extracted motion signal m is a partially missing signal, but since the motion information is mainly composed of low frequency components, if the missing signal is generated and used in the interpolation circuit 63, it can be used sufficiently. It is practical. after that.

A coefficient circuit 64 converts it into a motion control coefficient k. Note that the coefficient circuit 64 corresponds to the nonlinear conversion circuit 54 in FIG.

[Problem to be solved by the invention]

First, in the first reference document, P
Two methods are described that utilize the difference between one frame and the difference between four frames of the signal.

In either method, consideration is not given to sample formation11.

Therefore, when detecting motion based on the difference between each frame of the P signal, for a picture with large bichromaticity and rich in two hues, the multiplexed C component is not canceled out even in the case of a still image, and "movement is detected" is detected. This will generate a signal. Further, even if the characteristics of the spatial filter circuit 53 shown in FIG. 5 are restricted to suppress the color band components, the motion information of the two color components will not be detected.

In the first order, where motion detection is missed, motion detection is performed using the difference between four frames of the P signal, and since the phases of the C signal have the same polarity, there is no possibility of mistaking stationary motion for motion. Since the temporal distance of the images to be calculated becomes long, in the case of fast-moving moving images, motion detection may be missed. Therefore, no matter which motion detection method is adopted, that is, the difference between 1 frame or the difference between 4 frames, the accuracy of motion detection is poor, resulting in errors in the signal processing form of YC separation, resulting in poor image quality. There was a problem of deterioration.

On the other hand, the motion detection method based on the difference between one frame described in the second reference document places restrictions on the sampling frequency and its phase when sampling the P signal, so This differs from the motion detection method based on frame-to-frame differences described in the reference literature.

However, erroneous motion detection due to multiplexing of C signals does not occur.

For the YC separation format of 2 frames and 2 lines, omissions occur in motion detection. For example, as shown in Figure 7,
Figure (a) shows a rectangular object image that gradually becomes brighter from right to left over the course of two hours.

Assuming a moving image that moves horizontally from right to left in the order of (b) and (C), first, movement information is detected from the difference between the current frame and the image one frame before, so in the same figure As shown in (d), the moving region is the portion α shown in a grid pattern, and β is the stationary region. Therefore, in the still region β of the motion information (d), YC separation is processed by calculation between two frames. However, regarding YC separation, the still area β of the motion information (d) must essentially be a moving area. This clearly indicates a failure to detect motion and an error in signal processing for YC separation. As a result, there was a problem in that the image two frames before was displayed as a ghost on the monitor, deteriorating the image quality.

The purpose of the present invention is to solve the above-mentioned problems in the prior art.
By optimizing the YC separation signal processing for image movement,
Eliminate false detection and omission of motion.

An object of the present invention is to provide a motion detection circuit for PAL color television signals that can improve motion detection accuracy.

PAL color television signal at four times the color subcarrier frequency,
and a first motion signal sampled with a sampling frequency signal that matches the phase of the color burst signal and detected from a difference signal between the current frame and the previous frame of the sampled signal; The circuit configuration is such that a motion signal is obtained by synthesizing a second motion signal detected from a low frequency component of a difference signal between the current frame and two previous frames.

[Effect]

If the P signal is sampled at four times the color subcarrier frequency and at a sampling frequency that matches the phase of the color burst signal, by creating a difference between one frame of the sampled signal,
It is possible to target the movement of the C signal component and the movement of the Y signal component, and moreover, it is possible to detect fast movements. Furthermore, the motion signal detected from the low frequency component due to the difference between two frames of the sampled P signal compensates for the motion that was not detected due to the difference between two frames. Thereby.

This prevents omission of motion detection in determining a moving image as a still image, and prevents mistakes in the signal processing form of YC separation related to motion detection.

〔Example〕

Two embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 shows an embodiment of a YC separation circuit for PAL color television signals according to the present invention.

The P signal from the input terminal 1 is sampled by a sampling circuit 2 including an A/D (analog-to-digital) conversion function. The clock signal for sampling here is generated by a clock generation circuit 3 having a phase adjustment function, and is set to four times the frequency fsc of the two-color subcarrier (4fsc) and a phase that matches the phase of the color burst signal. be done. A portion of the sampled P signal is sent to the first line memory circuit 4.
2HD and 2FD (however, HD is a horizontal scanning period,
D represents the frame period, subject to a delay of FD=625HD). Arithmetic circuits 8 and 9 are subtraction circuits for producing a difference signal CL between two lines and a difference signal CF between two frames, respectively.

and a difference signal CL between two lines generated by the arithmetic circuit 8.

The difference signals CF between two frames produced by the arithmetic circuit 9 are both carrier color signals and have different characteristics depending on the nine images. If the CL multiplied signal and CF signal are selected incorrectly, the YC separation characteristics will deteriorate, so it is necessary to properly use the C signal and the CF signal depending on the two images. In this embodiment, the CL multiplied signal CF signal is synthesized in a mixing circuit 10 using a motion control coefficient k generated by a motion detection circuit 20, which will be described later, in a quadratic relationship, and the high This is to obtain a high quality C signal.

C=k ICL+ (1-k)·Cp However, 0≦ and ≦1. That is, for moving images, the motion control coefficient k becomes large, and the ratio of Ct and signal components to the synthesized C signal increases.On the other hand, for pictures from still images, the motion control coefficient k approaches O, and the CF signal The proportion of

Next, the C signal synthesized in the mixing circuit 10 is passed through a color bandpass filter 11 centered at the frequency fsc of the two-color subcarrier to remove the DC component and obtain the desired C signal.
It becomes a signal. This C signal is returned to an analog signal using a D/A (digital-to-analog) conversion circuit 14 and outputted to an output terminal 15 . Furthermore, since the P signal is a combination of the Y signal and the C signal, the sampled P signal is sent to the delay circuit 12.
If the C signal (output signal of the bandpass filter 11) extracted by the signal processing described above is subtracted by the arithmetic circuit 13, a high-quality Y signal that has been adaptively processed for motion will be obtained. Using this Y signal, the D/A conversion circuit 16
The signal is converted back to an analog signal and output to the output terminal 17.

The operation of the motion detection circuit 20 will be explained in more detail using FIGS. 2 and 3. First, the principle of motion extraction of the P signal will be described. In Figure 2, (F-1),
and FIl indicate the waveforms of the modulated color difference component U and V signals in the nth line of two consecutive frames. The solid line is the U signal, and the two-dot line is the V signal. In addition, when the S signal is sampled from the P signal at a frequency of 4fac and a clock whose phase matches that of the color burst signal, Fn
and CF-1)n, that is, the difference signal between one frame, for five consecutive pixels. In the same figure, the sample values of the C signals of five consecutive pixels (F-1) are CO= (uo+va)/
v'Y C,=: (ut-vJ/sf"I C2= (-u,-vo)/v" Ca== (-ua+Vs)/a c,= (u*+v4)/vT. Also , the sample value of the C signal of Fll is co = (
ua-vo)/sf'l= C engineering= (u, -vl)/a Cz=(u2+V2)/s/"X C3= (-u3+va)/v'W C4= (-u4-VJ/a So, based on the sample value of the above C signal.

When calculating the difference signal S between Fn and (F-1)ll, the first pixel:
5o=v'! (-u-va) 2nd pixel: S = 0 3rd pixel: 5z = v'T (u2 + vJ 4th pixel: S,
=O 6th pixel: S*=v'E'(-u4-v,) If it is a still image, the signals of pixels with the same polarity relationship (2nd pixel and 4th pixel) are canceled out and O becomes. However, in the case of a moving image, a significant (motion) signal is generated here. Pixels having the same polarity as described above have a period of one pixel.

As shown in FIG. If you do it.

An S signal containing a motion signal for every other pixel can be obtained. This S signal is sent to the sample circuit 22 by 1/(2
fsc).

Only the motion signal m can be obtained. However.

The amount of motion information of this m signal is thinned out every other pixel with respect to the amount of motion information of the video signal. Therefore, it is necessary to supplement the motion signal m that has been thinned out by two through interpolation processing. In this embodiment, before interpolation processing, the m signal is first input to the absolute value circuit 23 and converted to an absolute value signal of 1 ml, thereby forming a signal representing the magnitude of movement. after that,
Interpolation processing is performed by applying 1 ml of the absolute value signal to the interpolation circuit 24, and a motion signal corresponding to the pixel unit of the video signal is obtained.

It is converted into a motion signal having a so-called sampling frequency of 4 fsc. This is the first value obtained by the difference between one frame.
This is a motion signal M engineer. Even if the interpolation signal at this time is substituted with the left and right average values of the decimated 1 ml signal, the interpolation accuracy will not be impaired, and the average value interpolation can be easily realized with a circuit configuration.

On the other hand, the m P signals are delayed by two frame periods via the first frame memory circuit 6 and the second frame memory circuit 7, and the arithmetic circuit 9 performs subtraction with the P signal without delay. As mentioned above, by doing this, a CF multiplied signal can be obtained for a static color image. However, when the video is moving, a motion signal is mixed in with the CF multiplied signal. Therefore, a part of the output signal of the arithmetic circuit 9 is applied to the low-pass filter 25 to remove the signal of the two-color band reduction component, thereby extracting only the motion signal. Furthermore, in order to obtain a signal of the magnitude of the motion signal, the absolute value circuit 26 converts it into an absolute value signal.

This is the second motion signal M2 obtained by the difference between two frames.

The first motion signal M1 and second motion signal M2 formed as described above are combined in the combining circuit 27 to form a new motion signal M. As a method of synthesis, 2. For example, 1st
By outputting the larger value of the second motion signal M and the second motion signal M2, it is possible to compensate for each other's undetected motions.

Since the motion signal M formed by the synthesis circuit 27 has a low signal level, it is amplified and subjected to nonlinear processing in the nonlinear conversion circuit 28. For example, there is a method using nonlinear characteristics as shown in FIG. This figure is an example of converting the output level to 4-bit precision (16 levels) when the input motion size is 8-bit precision (256 levels).

The motion signal processed by the nonlinear conversion circuit 28 is expanded in space and time by a two-time-spatial integration circuit 29, and outputted as a motion control coefficient. By extending the motion into space and time, it is possible to correct the failure to detect motion.

As explained above, according to the circuit of this embodiment, the first motion signal M is formed by the difference signal between one frame of the signal obtained by sampling the P signal with a predetermined sampling signal. The second motion signal M2, which is effective for detecting fast motion and is formed by the difference signal between two frames of the 111 P signal, is detected by the difference signal between one frame. This is effective in supplementing missing movement information. The motion control coefficient includes
Since the first motion signal Mi and the second motion signal M2 are synthesized, it is possible to significantly improve motion detection accuracy.

〔Effect of the invention〕

According to the present invention, in motion detection of a PAL composite color television signal, failure to detect motion such as determining a moving region as a still region is improved. By performing yc total separation, etc., the image quality of the moving region of the video can be significantly improved.

[Brief explanation of the drawing]

Figure 1 is a circuit diagram showing an embodiment of the present invention, Figure 2 is a circuit diagram showing an embodiment of the present invention.
FIG. 3 is a diagram showing a color difference signal and an interframe difference signal of an L color television signal, and FIG. 3 is a diagram showing an example of nonlinear conversion characteristics in the motion detection circuit of the present invention. FIG. 4 and FIG. 5 are circuit diagrams explaining the conventional technology, respectively;
FIG. 6(a) is a diagram showing an example of motion detection in the prior art, and FIG. 6(b)
FIG. 7 is a diagram showing a difference signal between one frame in FIG. <Explanation of symbols> 1...Input terminal 2...Sampling circuit 3...
・Clock generation circuit 4, 5...1 line memory 6.
7...1 frame memory 8.9, 13.21...Subtraction circuit 10...Mixing circuit 11...Band pass filter 12...Delay circuit 14.16...D/A conversion circuit 15.17 ...Output terminal 20...Motion detection circuit 22
...Sample circuit 23.26...Absolute value circuit 2
4...Interpolation circuit 25...Low pass filter 27...Synthesizing circuit 28...Nonlinear conversion circuit 29...Spatio-temporal integration circuit M...Motion signal P...
PAL color television signal Y...Brightness signal C...Carrier color signal k...
・Motion control 6 u, v...Modulated color difference signal Attorney Junnosuke Nakamura (o) 2 frames per night (c) i-frame TV screen (d) recommendation! Seichitsu wisiiro

Claims (1)

[Claims] 1. In a motion detection circuit that separates a luminance signal and a carrier color signal from a PAL composite color television signal according to image movement, the PAL color television signal is divided into four times the color subcarrier frequency. and a means for sampling with a sampling signal that matches the phase of the color burst signal, and detecting from a difference signal between the current frame and the previous frame of the sampled signal. motion signal, current frame and 2
1. A motion detection circuit for a PAL color television signal, characterized in that a second motion signal detected from a low frequency component of a difference signal between a previous frame and a second motion signal is synthesized to obtain a motion signal. 2. A motion detection circuit for a PAL color television signal, characterized in that the larger one of the first motion signal and the second motion signal according to claim 1 is selected as the motion signal.