JPH01183991A - Movement detecting circuit for chrominance signal - Google Patents

Abstract

PURPOSE:To obtain a good chrominance movement signal without detection omission or error detection by using a difference signal between two frames as the chrominance movement signal and averaging it with the aid of a low pass filter besides. CONSTITUTION:A frame difference signal detection circuit 27 is constituted by a frame memory 25 having the delay time of one frame and a subtracter taking the difference between the input and output signal of it. The frame difference signal of an inputted television signal is detected and supplied to a band-pass filter 43. Then, the chrominance signal included to the first frame difference signal is extracted and separated. The two frames difference signal is made to be an absolute value by an absolute value circuit 47 and smoothed by the low-pass filter 48 after that. Thereby, a signal corresponding to movement information between the two frames of the chrominance signal is obtained. Thus, a movement detection signal (r) is made to be zero when it is a still picture. In case of an image of which contour of a substance to be filmed is the same but only a color component in the contour is different, the substance to be filmed can be judged as a moving picture.

Description

DETAILED DESCRIPTION OF THE INVENTION "Field of Industrial Application" This invention is a color signal motion detection circuit suitable for application to high-definition television receivers such as IDTV, and in particular, a color signal motion detection circuit using a field memory. The present invention relates to a color signal motion detection circuit that improves image quality by performing calculations in the axial direction.

``Prior art'' When reproducing images sequentially using interlaced scanning, as in current television receivers, interference such as line flicker, such as flickering of fine parts of characters, and interference between luminance and color signals occur. disturbances due to incomplete separation;
That is, dot interference, cross color, etc. occur.

In order to prevent such image deterioration, high-definition television receivers have been developed.

As is well known, this high-definition system displays images by converting the current television signal (interlaced signal) into a non-interlaced version.

To do this, create an interpolation scanline between the current scanlines,
It is necessary to convert interlaced scanning to non-interlaced scanning. This reduces line flicker.

Furthermore, by performing arithmetic processing between frames, incomplete separation of the luminance signal and color signal is removed, and dot interference and cross color are removed.

These processes will improve image quality.

Now, in creating the above-mentioned interpolated scanning line, it is sufficient to create an interpolation signal from the preceding and following fields in a still image.

However, in the case of videos, there is a time lag between fields, so creating an interpolated signal from the previous and subsequent fields may result in double images after non-interlacing.
Alternatively, the outline may be hidden.

Therefore, in the case of a moving image, it is sufficient to stop creating interpolation signals from the preceding and following fields, and instead create interpolation signals from the upper and lower lines within the field.

Similarly, when it comes to signal separation between luminance and chrominance signals, in the case of moving images, accurate separation is impossible due to the time lag, so signal separation is performed within the field as in the past. do it.

Therefore, in such a high-definition television receiver, it is necessary to switch the signal processing means according to the movement of the input subject. Therefore, such high-definition television receivers are equipped with a circuit for detecting motion information of a human-powered subject, that is, a motion detection circuit.

FIG. 10 is a system diagram showing an example of a conventional motion detection circuit 10.

In FIG. 10, a television signal, which is an input signal, supplied to terminal 1 is supplied to a delay means comprised of a plurality of memories, and each is delayed by a predetermined time.

In this example, there are three field memories 2.4°5 and 2.4°5.
Two line memories 3 and 6 are connected in cascade as shown in the figure.

If point b of the connection is taken as a reference, in the subtracter 7,
The difference between the respective television signals obtained from point b and point d, which is one frame away from point b (one frame difference signal), is detected.

The level of the l-frame difference signal is reduced to 1/2 by an attenuator 8, and then a low-frequency luminance component is extracted by a low-pass filter 9, and the extracted low-frequency luminance component is passed to an absolute value circuit 1.
It is converted into an absolute value at 1.

Therefore, the absolute value circuit 11 detects a signal corresponding to the movement of the human-powered subject between frames. This motion detection signal is referred to as a first motion detection signal.

Similarly, the television signals obtained from point a and point 0, which are fields before and after point b, are supplied to the subtracter 12, and a one-frame difference signal is detected in the same manner as described above.

This frame difference signal is supplied to a low-pass filter 14 via an attenuator 13, and only low-frequency luminance components are extracted and separated.
The output is further converted into an absolute value in an absolute value circuit [5].

If the output converted to absolute 1 is used as the second motion detection signal,
These motion detection signals are further supplied to an adder circuit 16 and combined.

Therefore, a motion detection signal corresponding to the movement of the human-powered subject is obtained at the output terminal 17.

The pixel arrangement of each point a + b + c + d is as shown in the 110th pixel arrangement. Figure A shows a case where time is taken as the composition axis and the vertical axis is taken in the vertical direction. If this is expressed in two dimensions (horizontal and vertical), it will be as shown in figure B.

Now, in the case of a still image, it is thought that the same signal is coming from one frame to the next, so the difference is 0. In the case of a moving image, there is a shift in the image between frames due to movement, so the shifted part some change occurs. Therefore, the movement of the subject can be detected by taking the difference between frames.

Therefore, when the above-mentioned motion detection signal is obtained,
It is determined that the image is a moving image, and the signal processing means is switched and controlled accordingly.

Here, in addition to the motion detection signal detected from the difference between the reference point S and point d, which is one frame away, the motion detection signal detected from the frame difference between fields a and C before and after the reference point S is used as follows. Based on the following reasons.

That is, in interpolation processing when converting interlaced scanning to non-interlaced scanning, television signals of fields before and after the reference point are used as interpolation data for still images.

Therefore, it is necessary to detect field movements before and after the reference point. Note that the motion information detection range at this time is the area surrounded by the solid line in FIG.

By the way, when configured in this way, the following problems arise.

That is, in the above-described configuration, since the motion detection circuit 10 is provided with the low-pass filters 9 and 14, the motion information of the input subject is only the motion information of the low-frequency luminance component obtained from the human-powered subject.

Therefore, even if the image is a still image, the movement of the human-powered subject cannot be accurately detected, especially in the case of a human-powered subject where only the color signal (color component) changes.

An example of an object that is a still image and differs only in color components is, for example, an image created using computer graphics.

This is because, in the field of computer graphics, for example, drawing techniques that express the same contour in different colors are often employed.

Therefore, when targeting such subjects,
As a result of being judged as a still image, interpolation processing is performed using adjacent field signals, resulting in deterioration of image quality after interpolation.

In order to solve such problems, some conventional motion detection circuits 10 are configured to detect motion of color signals in addition to motion detection mainly based on the luminance component described above.

FIG. 12 is a system diagram showing an example of this, and the diagram shows only the portion that detects the movement of the color signal.

The input television signal supplied to the terminal 1 is delayed by ] frames in the frame memory 25, and the delayed signal is supplied to the bandpass filter 71, whereby only the carrier color signal, that is, the color component, is extracted from the delayed television signal. Extracted and separated.

The extracted color components are further processed in two east circuits 72.
The product is multiplied. In the squaring operation, the phase of the color component is inverted every frame, so this circuit is used to prevent this phase inversion.

The squared color signal is supplied to a subtracter 73.

On the other hand, the input television signal supplied to the terminal 1 is further supplied to a bandpass filter 74, where a color signal in the same frequency band as mentioned above is extracted and separated, and a 2-east circuit 75 at the subsequent stage performs a squaring operation. be done.

Then, each square calculation output is subtracted by the subtracter 73, whereby a difference signal between one frame is detected. The one-frame difference signal is unchanged, and serves as motion information of color components between frames.

This motion detection signal is converted into an absolute value in an absolute value circuit 47.

By the way, when the color signal motion detection port ¥a40 is configured in this way, the frame difference signal of this color signal is used as the motion detection signal, so for example, the above-mentioned This may cause inconvenient problems during interpolation processing in computer graphics.

In computer graphics, etc.
Image creation processing may be performed such that the phase of the color signal is in the same phase between frames, so the movement of the color signal in such a case is different from that of a still image where the phase of the color components is reversed. They become indistinguishable.

Therefore, in such a case, there is a drawback that the motion detection is inaccurate.

Strictly speaking, the bandpass filters 71 and 74 extract high frequency components of the luminance signal in addition to the color signal.

Therefore, an erroneous detection occurs in a portion where a high-frequency component of a luminance signal exists, such as an outline portion of an image, even though it is a still image.

Therefore, the present invention proposes a color signal motion detection circuit that improves the above-mentioned drawbacks of the conventional circuit and makes it possible to obtain more accurate color motion information.

That is, this invention focuses on the fact that the phase of the color signal is inverted for each frame, and that the inverted phase is further inverted between two frames to become the same phase. We also try to detect the movement information of color signals by taking .

It is.

"Means for Solving the Problems" In order to solve the above-mentioned problems, the present invention provides motion detection of color signals of a television receiver that reproduces television images by interlace scanning color television signals. In the circuit, multiple field memories and
It is equipped with an addition/subtraction circuit, a bandpass filter that passes frequencies centered around the color subcarrier frequency, and an absolute value conversion circuit, and color components are extracted by manually inputting the difference signal between one frame to the bandpass filter. At the same time, by summing the input and output signals having a one frame difference in the second frame memory, a two frame difference signal regarding the color component is detected, and this two frame difference signal is used as a motion detection signal regarding the color component. It is characterized by being made to

"Operation" In this configuration, the frame difference signal detection circuit 27 detects the difference No. 18 between one frame of the human-powered television signal.

An inter-frame sum signal of the frame difference signal is detected in a circuit 46 (hereinafter referred to as a frame sum signal detection circuit) which roughly calculates the sum between frames. Therefore, the frame sum signal detection circuit 4G detects the difference signal between two frames of the human-powered television signal.

In this way, if the output obtained by detecting the difference signal between two frames is used as the motion signal of the color signal, it is possible to reliably detect the motion of only the color component as described above. .

Furthermore, since this color signal detection system is provided with a frame difference signal detection circuit 27, in a still image, the luminance signals between adjacent frames are completely canceled out or almost zero. Become.

Therefore, the high-frequency components of the luminance signal remain and are not output from the bandpass filter 43. As a result, there is no possibility of false detection caused by high-frequency components of the luminance signal.

This allows a change in the color signal to be detected.

``Example'' Next, details will be given with reference to FIG. 1 and subsequent figures regarding the case where an example of the color signal motion detection circuit according to the present invention is applied to the motion detection circuit of the above-mentioned high-definition television receiver. Explain.

FIG. 1 is a system diagram showing an example of a color signal motion detection circuit 40 according to the present invention, in which an input television signal supplied to a terminal l is supplied to a frame difference signal detection circuit 27.

The frame difference signal detection circuit 27 includes a frame memory 25 having a delay time of one frame, and a subtracter 41 that takes the difference between the input and output signals. Ru.

In the case of a still picture, the luminance signals are in phase even between adjacent frames, so when the input television signal is supplied to this detection circuit 27, the luminance signals are completely canceled out.

After the frame difference signal is attenuated to a predetermined level by an attenuator 42, it is supplied to a bandpass filter 43, where the color signal included in the first frame difference signal is extracted and separated. The extracted and separated color signals are further sent to a frame sum signal detection circuit 46.
is supplied to

The frame sum signal detection circuit 46 includes a frame memory 44 and an adder 45 that sums its input and output signals.

As a result, the detection circuit 4G detects a difference signal spanning two frames.

Therefore, with this configuration, it is possible to obtain motion information equivalent to taking the difference between two frames and band-limiting the output.

This two-frame difference signal is converted into an absolute value by an absolute direct circuit 47, and then smoothed by a low-pass filter 4, thereby obtaining a signal corresponding to motion information between two frames of the color signal.

As described above, the motion detection signal related to the color signal is supplied to the interpolation processing switching circuit as its control signal.

With the configuration as described above, motion information equivalent to taking the difference between two frames and band-limiting the output can be obtained, but this will be explained using a mathematical formula.

Assume that the human signal is a discrete numerical sequence, and this is expressed as x(n)
(n=0.1.2...), attenuator 4
The output from 2 is 1/2 (x (n) - x (n - F) ) (F:
1 frame delay).

Letting the impulse response of the bandpass filter 43 be H(n), its output is 1/2(x(n)-x(n-F))ell(n)(Thursday:
convolutional sum). When this signal is input to the frame memory 44 and the sum of its input and output is calculated, 1/2[(x(n)-x(n-F))neII(n)+(
x(n-F)-x(n-2F)) Kone II (n)=
1/2(x(n)-x(n-2F)) zero It (n)
It can be seen that this is equivalent to processing that takes the difference between two frames and then passes them through a bandpass filter.

In the present invention, since the movement of the color signal is detected by taking the difference between two frames, the motion information may be obtained by directly taking the difference between the two frames.

In the embodiment, instead of using such means, the difference between one frame is calculated, and then the sum of the signals between each frame is calculated. This is based on the following reasons.

In other words, if the output signal of the bandpass filter 43 is amplitude-compressed and then input to the frame memory 44, the circuit scale of the frame memory 44 can be reduced compared to the previous frame memory 25.

Now, in the NTSC television signal, the phase of the subcarrier of the color signal is inverted for each frame, so the frame difference signal detection circuit 27 detects the n-1 frame and n-1 frame.
A frame difference signal p, which takes the difference between frames, and a frame difference signal q, which also takes the difference between frames n and n+1.
If the sum is further calculated in the frame sum signal detection circuit 46, different motion detection signals r are obtained when the color component is stationary and when it is moving.

That is, in the case of a still image, the frame difference signal p and the frame difference signal q have completely the same waveform. However, only the phase is reversed.

The frame difference signals p and q are detected by the frame sum signal detection circuit 46.
Therefore, in the case of the above example, the final motion detection signal r becomes zero.

In other words, it is determined to be a complete still image.

On the other hand, if only the color signal differs from frame to frame, the levels of frame difference signals p and q will differ. As a result, the detection signal r does not become zero.

In other words, for images in which the outline of the subject is the same but only the color components within the outline are different, the subject is determined to be a moving image, and the interpolation processing corresponding to the subject is determined to be a moving image.

Now, regarding the movement of color components, it is necessary to consider the following points.

That is, if the phase of the subcarrier is reversed between two frames due to movement of the color component, the second
As shown in the figure, although the object is moving in a certain area, a motion detection signal is output in which a motion output and a still output are repeated for each drawing.

In order to deal with this problem, as described above, a low-pass filter may be provided at the subsequent stage of the absolute value circuit 47 to average the motion detection signals.

As the low-pass filter 48, a low-pass filter having a configuration as shown in FIG. 3 can be used.

This low-pass filter 48 has a pair of cascade-connected delay elements 91 and 92, the human input signal and the final output signal are added in a first adder circuit 93, and the first stage delay element 91.
After the output delay number 1g is doubled in gain by the amplifier 94, it is added to the addition output of the first addition circuit 93 and the second addition circuit 95. The added output is attenuated to 1/2 by an attenuator 96 and output.

With this configuration, averaging is achieved between adjacent pixels, so if a motion detection signal like the one shown in Figure 4 is manually generated by the low-pass filter on the 4th day, an output like the one shown in Figure 4 will be generated. It is obtained as follows.

That is, since adjacent pixels are averaged and output, the above-mentioned inconvenient problem can be reliably solved.

By the way, the color signal motion detection circuit according to the present invention can of course be used alone, but as mentioned above, by using it together with the luminance signal motion detection circuit, the purpose of Hatsutomo can be achieved. It is something that can be demonstrated without any effort.

Such an embodiment will be described below.

As shown in FIG. 5, the human television signal supplied to terminal 1 is supplied to delay means 20. As shown in FIG.

As shown, the delay means 20 has three field memories 21, 23, 24 and one line memory 22 in this example.

These field memories 21+231 24 and line memory 22 are connected in series as shown in the figure.

Since a moving image is considered to have a certain degree of spatial area and temporal continuity, at least a two-dimensional area consisting of horizontal and vertical areas surrounding the reference point g (see Figure 9),
Preferably, movement is detected in three dimensional areas: horizontal, vertical and temporal.

The input/output signals at each stage of the delay means 20, that is, the human-powered television signal e and the predetermined delay signal f of the delay means 20
, h, i are supplied to the first detection circuit 30,
A one frame difference signal regarding the luminance component is detected. The delayed signals f and i are then supplied to a second detection circuit 40 to detect a two-frame difference signal regarding the color component.

The second frame difference signal detection circuit 40 is the color signal motion detection circuit according to the present invention described in FIG. For convenience of explanation, the motion detection signal of the color signal will be explained as a two-frame difference signal.

Both the l frame difference signal and the 2 frame difference signal are l
These are motion detection signals between frames and between two frames.

The 1-frame difference signal and the 2-frame difference signal are supplied to an adder circuit 51 at the subsequent stage and added. The motion detection signal 1, which is the output of the addition, is further supplied to a detection area enlarging circuit 60 to enlarge the motion detection range.

FIG. 6 and subsequent figures are configuration examples of each part of the motion detection circuit 10.

FIG. 6 shows an example of a one-frame difference signal detection circuit 30, which includes first and second one-frame difference signal detection circuits 30A and 30B.

Since both are for detecting a frame difference signal, in the first one-frame difference signal detection circuit 30A, the delayed signals e and h are supplied to the subtracter 31, and the first frame difference signal is detected. be done.

After this first frame difference signal is adjusted to a predetermined level by an attenuator 32, it is supplied to a low-pass filter 33 to extract motion information particularly regarding low-band luminance components.

The first frame difference signal is further converted into an absolute value in the absolute value detection circuit 34, thereby being converted to a level corresponding to the amount of movement of the human-powered subject.

The second one-frame difference signal detection circuit 30B uses the delayed signals f and i, thereby detecting the second one-frame difference signal.

The second one-frame difference signal detection circuit 30B is also configured in the same manner as the first one-frame difference signal detection circuit 30A, so a detailed explanation thereof will be omitted.

However, 36 is a subtracter, 37 is an attenuator, 38 is a low-pass filter, and 39 is an absolute value circuit.

In this way, the two one-frame difference signal detection circuits 30A,
30B to detect the motion of the low-range luminance component is that, as is clear from FIG. 9, motion information is also detected based on information from adjacent lines of adjacent frames. This is because the movement of the human-powered subject can be detected more accurately.

In FIG. 9, A shows the relationship between pixels in a plane perpendicular to the time axis. The black circle in the figure indicates the reference point on the time axis.

12II3 showed this using horizontal and vertical planes.

Although the one-frame difference signal detection signals obtained at the output terminals 35a and 35b are not shown, they are added in an adding circuit and used as the final motion detection signal j regarding the low-range luminance component.

In the embodiment, the motion detection area is further enlarged in order to avoid the motion detection of the input subject and improve the detection accuracy.

An example of this detection area enlarging circuit 60 is shown in FIG.

The motion detection signal l outputted from the addition circuit 51 in FIG. be done.

Furthermore, the detection area is expanded in the horizontal direction by the low-pass filter 65.

By doing so, the detection area is expanded in the horizontal and vertical directions (see FIG. 9).

As shown in FIG. 8, the frame memory 66 and the adder circuit 6
By providing a low-pass filter 68 in the time direction consisting of 7 and taking S of the input/output signal to the frame memory 66, expansion in the time direction can also be performed.

In this way, when the low-pass filter 6 days in the time axis direction is added, motion detection is performed in a range surrounding the reference point g in the three-dimensional area of horizontal, vertical, and time, as shown in Figure 9. can be done.

This is because the movement of an object is continuous in space and time to some extent, and the detection signal of that movement is also considered to be continuous in space and time.

In this way, when detecting the movement of a certain pixel, by taking into consideration the detection status of the surrounding pixels spatially and temporally, that is, by averaging the detection status of the surrounding pixels. , more accurate motion signal detection becomes possible.

When the detection area is expanded to include horizontal, vertical, and temporal three-dimensional areas, the detection range of motion information is as shown by the solid line in FIGS. 9A and 9B.

Note that the configuration of FIG. 1 and each component related thereto, and the detection range of motion information are merely examples, and the present invention is not limited thereto.

Of course, the detection area enlarging circuit 60 may only expand the detection area to a horizontal and vertical two-dimensional area.

"Effects of the Invention" As described above, in this invention, the difference between two frames is used as a color motion signal and is further averaged by a low-pass filter, thereby achieving good color motion without missing detections or such detections. I can get a signal.

For this reason, the present invention is extremely suitable for application to motion detection circuits used in high-definition television receivers such as IDTVs, as described above.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an example of a color signal motion detection circuit according to the present invention, FIGS. 2 and 4 are diagrams explaining its operation, and FIG. 3 is a system showing a specific example of a low-pass filter. Fig. 5 is a system diagram showing an example of a motion detection circuit that also includes a luminance component, Fig. 6 is a system diagram showing an example of a one-frame difference signal detection circuit, and Figs. 7 and 8 are respective detection areas. 9 is an explanatory diagram of the motion detection range. FIG. 10 is a system diagram of a conventional motion detection circuit.
FIG. 1 is an explanatory diagram of a conventional motion detection area, and FIG. 12 is a system diagram of a conventional color signal motion detection circuit. 1... Input terminal 10... Motion detection circuit 20... Delay means 30... 1 frame difference signal detection circuit 40... Color signal motion detection circuit (2 frame difference signal detection circuit) 60...・Detection area expansion circuit 27...Frame difference signal detection circuit 43φ...Band pass filter 46...Frame sum difference signal detection circuit 47...Absolute value circuit 48・◆Low bass filter patent applicant Sharp Corporation No. 30 48: cy-tI'27ft Mua 1st 4th Fig. 5 1i0: # Riki return praise g road Fig. 6 ≠ 2/717 mu 4 summer catch t3 ne 87 Fig. 6o, God and 9 door entry g protection Moe 8 Figure 9 A B ifE- Gatnis σ Zu Tora! Diamond Σ Fig. 10 1 Haku Nido Masu and Shosei No. 111'1

Claims (1)

[Claims] (1) The motion detection circuit of a television receiver that reproduces a television image by interlace scanning a color television signal includes multiple field memories, an addition/subtraction circuit, and a frequency detection circuit centered around the color subcarrier frequency. It is equipped with a band-pass filter for passing and an absolute value conversion circuit, and a color component is extracted by inputting a difference signal between one frame to the band-pass filter, and a one-frame difference is stored in a second frame memory. A color signal movement characterized in that a two-frame difference signal regarding a color component is detected by calculating the sum of input and output signals, and the two-frame difference signal is used as a motion detection signal regarding the color component. detection circuit.