JPH0373686A - Movement detecting circuit - Google Patents

Abstract

PURPOSE:To prevent a static image part even in a high frequency signal from being decided as a moving image part in error by providing the movement detecting circuit with a difference means for finding out a difference between the output signal of the 1st memory and a television signal inputted to the 1st memory and a means for processing the output signal of the difference means as a signal corresponding to a static image when the output signal of the 2nd memory satisfies a prescribed condition. CONSTITUTION:A digital comparator 33 converts a 2-frame difference signal 4 consisting of 4 bits into a 2-frame difference signal consisting of one bit which is turned to a high level '1' when the signal 4 exceeds its threshold and turned to a low level '0' when the signal 4 is less than the threshold and applies the converted 2-frame difference signal to a 2 frame memory 35 for applying the delay of a 2-frame period. Since the 2-frame difference signal and the delayed 2-frame signal are used, a moving image part accurately is detected, and even when an input signal is a high frequency signal exceeding 4MHz, a static image part is prevented from being decided as a moving image part in error.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is applicable to, for example, MUSE (Multiple 5u
b-Nyquist-5a+spring Encod
The present invention relates to a motion detection circuit suitable for use in a high-definition signal decoder of the ing) system.

[Summary of the invention]

The present invention is directed to a motion detection circuit for a television signal configured to construct one screen image using n-frame image signals, such as a MUSE high-definition signal. a first memory for providing a period delay; a difference means for determining the difference between the output signal of the first memory and the input television signal; and a difference means for giving a delay of n frame periods to the output signal of the difference means. A second memory is provided, and when the output signal of the second memory satisfies a predetermined condition, the output signal of the difference means is treated as a signal corresponding to a still image and processed. This prevents a still image portion from being erroneously determined to be a moving image portion not only for signals but also for high-frequency signals.

[Conventional technology]

Currently, it is considered practical to perform so-called high-definition broadcasting using broadcasting satellites. In this case, in order to transmit the high-definition signal as it is, a signal bandwidth of about 20 to 25 MHz is required, whereas the high-definition signal can be properly FM modulated and transmitted using one channel (bandwidth 27 MHz) of the broadcasting satellite. In order to do this, it is necessary to reduce the signal bandwidth of the high-definition signal to 9 MHz or less. Therefore, the MUSE system was developed to compress the bandwidth of high-definition signals and broadcast them on one satellite channel without degrading the image quality.

According to this MUSE method, the luminance signal Y and the color signal C are time-division multiplexed, and this time-division multiplexed signal (TCIC signal) is subjected to field offset, frame offset, or line offset subsampling (thinning), and the signal band The width is compressed to 8.1MHz. As a band compression method, so-called three-dimensional subsampling (thinning) is used in the MUSE method. That is, MUSE
As shown in Fig. 3, the high-definition signal output from the encoder has a phase difference of 180 degrees between the sample points of the first frame (third frame) and the sample points of the second frame (fourth frame). 1 by the 1st frame and the 2nd frame (or the 3rd frame and the 4th frame)
Complete image information for a screen is configured. Therefore, M
In the USE method, complete image information for one screen is constructed from signals for two frames (four fields).

However, a complete image for one screen can be constructed from the signals for two frames only when the original image is a still image.
When the original image is a moving image, if one screen's worth of images is constructed from the signals of two frames, double images or the like may occur in the reproduced image. Therefore, in the MUSE encoder, as is well known, a still image band compression circuit and a moving image band compression circuit are provided, which detect the amount of movement of the original image for each pixel and adjust the amount of movement according to the detected amount of movement. The output signals of the still image and video band compression circuits are mixed at a predetermined mixing ratio and supplied to the transmitter.In this case, the encoder side detects the one-frame difference between the current frame and the previous frame before subsampling. While motion can be easily detected using a signal, as shown in Figure 3, the signal input to the MUSE decoder side does not have a signal at the same sample point one frame before, making motion detection difficult. Become.

Output signal of the above MUSE encoder (MUSE signal)
The power to restore a high-definition signal with a bandwidth of about 20 MI (z) < MUSE decoder. Figure 4 shows the conventional MUSE decoder.
The main parts of the USE decoder are shown in Fig. 4 (
1) is an input terminal, and the input terminal (1) is supplied with a MUSH signal MS that has been FM demodulated in the receiving section and then analog/digital converted at 16.2 MHz.

This MUSE signal MS is supplied to a motion detection circuit (2), a still image interpolation circuit (3), and a moving image interpolation circuit (4), and the still image interpolation circuit (3) uses signals for two frames. Constructing a signal corresponding to one screen of images by interframe interpolation and interfield interpolation and supplying it to the mixing circuit (5),
The moving image interpolation circuit (4) performs intra-field interpolation from the signals for one field to form signals corresponding to one image for each field and supply them to the mixing circuit (5). In the mixing circuit (5), the output signals of the interpolation circuits (3) and (4) are mixed at a mixing ratio according to the motion detection signal KT supplied from the motion detection circuit (2), and are time-division multiplexed ( T.C.
I) supplied to the decoder (6).

In the motion detection circuit (2), the MUSE signal M S
o is the input terminal of frame memory (7), subtraction circuit (8)
and one input terminal of the subtraction circuit (9), and is the output signal of the frame memory (7).
The one-frame delayed signal MS is supplied to the input terminal of the frame memory (10) and the other input terminal of the subtraction circuit (8), and the two-frame delayed signal MS, which is the output signal of the frame memory (10), is supplied to the input terminal of the frame memory (10) and the other input terminal of the subtraction circuit (8). The output signal of the subtraction circuit (9) is supplied to the other input terminal of the subtraction circuit (9), and the output signal of the subtraction circuit (9) is supplied to the absolute value circuit (11
) as a two-frame differential signal Δ2
(12) is supplied to one input terminal.

Also, the output signal of the subtraction circuit (8) has a cutoff frequency of 4
It is converted into a delayed one-frame difference signal Δ1 via a MHz low-pass filter circuit (13), an absolute value circuit (14), and a frame memory (15), and this delayed one-frame difference signal Δ1 is sent to an inverter circuit (16). The signal is supplied to the other input terminal of the AND game 1-(12) through the signal.

In this case, as explained earlier with reference to FIG.
In the E signal, there is a signal at the same sample point two frames ago, but there is no signal at the same sample point l frame ago, so the two-frame difference signal Δ2 is a signal that corresponds to the amount of motion. , the delayed one-frame difference signal Δ1 does not completely correspond to the amount of motion. However, even if the sample points do not match, if the low frequency component is extracted, it can be considered that the sample points substantially match, and Mtl
In the case of SEt, the difference between one frame can be accurately determined for low-frequency signals of 4 MHz or less. In the example in FIG. 4, a low-pass filter circuit (13) is provided, so the delayed one-frame difference signal Δ1 is 4
The low frequency signal under MH2M is a signal that accurately corresponds to the amount of motion.

In addition, the 0 in the output signal of the AND gate (12) is delayed by one field each in the field memories (17), (18), and (19), and becomes a delayed signal.
, the output signal is 0 and the delayed signal is +,
Kz, and 3 are respectively supplied to an arithmetic circuit (20), and the motion detection signal KT is generated in this arithmetic circuit (20) by, for example, an OR operation. In the above circuit, each signal is represented as a 1-bit digital signal for simplicity, but in reality, the MUSE signal MS is, for example, 8 bits.

The motion detection signal KT is, for example, a 4-bit digital signal.

In the example in Figure 4, the MUS supplied to the input terminal (1)
To explain the operation when the E signal MS is a low-frequency signal of 4 MHz or less, the signal MS o is a high-vigilance signal related to an object moving at a constant speed, and the signal MS o is a high-vigilance signal related to an object moving at a constant speed. signal at position
Consider it as a function of

In this case, MUSE signal MS, 1 frame delayed signal M
S, and two-frame delayed signal MS2 are shown in FIGS. 5A and 5B, respectively.
and C, the delayed one-frame difference signal Δr
(lMSzMS+l) and the two-frame difference signal Δz(lMSz-MSII) are as shown in the fifth diagram and the previous figure, respectively. Therefore, the AND gate (
12> is the logical product of the signal Δ2 and the signal Δ, as shown in FIG. 5F.

What is important about this output signal being 0 is that the high level pulse (21) at position rZP+ is erased by the delayed one frame difference signal Δ1, leaving only the high level pulse at position P2. This means that when an object moves from position P+ to position Pt, only position P2 is determined to be a moving image part, and position Pl is determined to be a still image part.

Similarly, the output signal is sequentially delayed by l fields, and the delayed signals are as shown in FIG. 5 G-I, and are finally output from the motion detection circuit (2). The motion detection signal KT is as shown in FIG. 5J. In FIG. 5J, the area (22) is determined to be a moving image portion, and the area (23) is accurately determined to be a still image portion. In addition, 4
The motion detection signal K is calculated using 0 to 0 for the signal for one field.
The reason why T is synthesized is because the still image interpolation circuit (3) of the MUSE decoder composes a complete image for one screen from signals for two frames (four fields). 1 field even if it is determined to be a still image part
This is because the portion determined to be a moving image portion in the field three fields before must be determined to be a moving image portion.

[Problem to be solved by the invention]

As mentioned above, the example of FIG. 4 can accurately perform motion detection for low frequency signals of 4 MHz or less.

However, in the conventional motion detection circuit (2), the delayed one-frame difference signal Δ does not respond to high-frequency signals exceeding 4 MHz, and therefore, regarding high-frequency signals exceeding 4 MHz, still image portions are mistakenly treated as moving image portions. There was an inconvenience that there was a risk of judgment.

That is, in the example of FIG. 4, when the MUSE signal M So supplied to the input terminal (1) is a high frequency signal exceeding 4 MHz, the signals of each part become as shown in FIG.
As is clear from the output signal 0 in Figure F, the object is at position P+
Even when moving from to position Pt, position Ml p
The pulse (24) at + remains as is. Similarly, since the pulses from two frames before remain in the delayed signals 1~ and (Fig. 6G~■), the final motion detection signal does not match the final motion detection signal (Fig. 6J). In addition to the original video part (25), a high-level pulse of 1' is generated in the part (26) that is originally a still image, and it is determined to be a video part. If the part that is a picture is determined to be a moving picture part, the mixing circuit shown in Figure 4 (
Since the output signal of the moving image interpolation circuit (4) is selected in step 5), there is an inconvenience that the resulting image may be partially blurred.

Note that various motion detection circuits have been proposed for the normal NTSC-M system (for example, Japanese Patent Laid-Open No. 63
-90988, JP-A-62-175093.

Utility Model Publication No. 62-203583, Japanese Patent Application Publication No. 61-708
(Refer to Publication No. 88), no other suitable system has been proposed for the high-definition signal of the MUSE method.

In view of the above, the present invention provides a motion detection circuit for a television signal that uses multiple frames of image signals, such as MUSE high-definition signals, to construct one screen image. To prevent a still image portion from being mistakenly determined as a moving image portion even for high-frequency signals.

[Means to solve the problem]

The motion detection circuit according to the present invention is, for example, as shown in FIG. , a first memory (28) that delays the input television signal M S o by n frame periods; an output signal MS of the first memory (28); and the input television signal MS. , and a second memory (35) that delays the output signal (Δ2, Δ4, etc.) of this differential means (29) by n frame periods. When the output signal Δ, of the memory (35) of the memory (35) satisfies a predetermined condition (for example, exceeds a predetermined level), the difference means (
The output signal (Δ□Δ4, etc.) of 29) is treated as a signal corresponding to a still image and processed (for example, the output signal is forcibly set to zero level).

[Effect]

According to the present invention, first, a primitive signal corresponding to movement is obtained by the difference means (29). With this primitive signal, for example, when an object moves from position PI to position Pt, not only position P2 but also portion P, which was originally a still image portion, is determined to be a moving image portion. Therefore, in the present invention, when the output signal Δ3 of the second memory (35) satisfies a predetermined condition, the difference means (2
By processing the primitive signal of 9) as a signal corresponding to a still image, it is possible to prevent a still image portion from being erroneously determined to be a moving image portion.

At this time, since the present invention does not use a low-pass filter circuit, the moving image portion can be accurately detected not only for low-pass signals but also for high-frequency signals.

〔Example〕

Hereinafter, the first embodiment of the motion detection circuit according to the present invention will be described.
This will be explained with reference to FIG. This example is a motion detection circuit of a MUSE decoder (i.e., when n=2 in the present invention)
In FIGS. 1 and 2, parts and signals corresponding to those in FIGS. 4 to 6 are denoted by the same reference numerals, and detailed explanation thereof will be omitted.

FIG. 1 shows the motion detection circuit of this example. In FIG. 1, (27) indicates an input buffer, and 16. Made by analog/digital conversion at 2MHz,
The input port of the 2-frame memory (28) and the subtraction circuit (
29), and the 2-frame delayed signal MS which is the output signal of this 2-frame memory (28).
, is supplied to the other input port of the subtraction circuit (29), and the m2-bit output signal (MS
o MSz) to the absolute value circuit (30). This absolute value circuit (30) converts its output signals (MS, -MS) into absolute values, then extracts the upper m3 bits of the absolute values and generates a two-frame difference signal Δz (=l MSo
MSz l), and this two-frame difference signal Δ2
is supplied to the sensitivity setting device (31). This sensitivity setting device (3
1) converts the m-bit difference signal Δ2 into a 4-bit 2-frame difference signal Δ4. In this example, m1=8. mz =9. It can be set to m3=6. m
As a method for converting the 3-bit difference signal Δ2 into the 4-bit difference signal Δ4, in addition to the method of extracting the upper 4 bits of m3 bits, logarithmic compression or the like may be used.

The signals of each bit of the difference signal Δ4 obtained by the sensitivity setter (31) are connected to AND gates (32A) to (32D), respectively.
), and at the same time its difference signal Δ4
is supplied to the non-inverting input port of the digital comparator (33), and 4-bit reference data corresponding to a predetermined dark value is supplied from the level setter (34) to the inverting input port of the digital comparator (33). .That is, this digital comparator (3
3) is a 1-bit signal that becomes high level "1n" when the 4-bit 2-frame l signal Δ4 exceeds its dark value and becomes low level "0" when the difference signal Δ4 is less than its dark value. This l-bit 2-frame differential signal is converted into a 2-frame differential signal and is supplied to a 2-frame memory (35) for providing a delay of 2-frame periods. 1-bit delay on output port 2
The frame difference signal Δ3 is commonly supplied to the other input terminal of each of the AND gates (32A) to (32D) via an inverter circuit (36).

The field memory (
17), (18), and (19) sequentially add a delay time for one field period, and the 0 and 4-bit delay signals K + "' K3 are respectively applied to the 4-bit output signal by the arithmetic circuit ( 20), and the arithmetic circuit (20) receives, for example, a signal K.

A 4-bit motion detection signal KT is generated by performing an OR operation on each of the 3 bits of .about., and this motion detection signal KT is supplied to the mixing circuit (5) in the example of FIG.

As explained earlier with reference to FIG. 3, in the MUSE system high-definition signal, there are no signals at the same sample point between frames, so only motion detection signals for low-frequency signals can be obtained. Since there are signals at the same sample point within one frame, accurate motion detection signals can be obtained not only for low-frequency signals but also for high-frequency signals. That is, the 4-bit 2-frame difference signal Δ4 output from the sensitivity setting device (31) in the example in FIG.
All of the high-frequency signals exceeding 200 kHz accurately contain motion information. Therefore, the delayed 2-frame difference signal Δ obtained by compressing the 4-bit signal Δ4 to 1 bit and adding a delay time of 2 frame periods is also accurate for both the low-frequency signal and the high-frequency signal. contains movement information.

In the example of FIG. 1, the operation when the MUSE signal M S o supplied to the input buffer (27) is a high-definition signal related to an object moving at a constant speed will be described with reference to FIG. 2. The value of each signal is displayed as a 1-bit digital signal, and each signal on the current horizontal scanning line is displayed as a function of position X. Further, in this example, the MUSE signal MS operates in common regardless of whether it is a low frequency signal or a high frequency signal.

In this case, MtlSE signal MS, , 2-frame delayed signal MS□ delayed 2-frame difference signal Δ3 (=1MS4-M
S1, MS4 are 4-frame delayed signals) and 2-frame differential signals (=lMS, -MSOL) are shown in FIG. 2A-, respectively.
As shown in FIG. 2, the signal that is the logical product of the signal Δ4 and the signal τ is 0, as shown in FIG.
The inverted signal Δ.

Due to the action of , a high level “1” pulse at position P+ (
37) has been deleted. Therefore, the position WPz
Since only the high level “1” pulse remains,
As shown in FIG. 2 F-H, unnecessary pulses from two frames before the delayed signal and ~ are respectively erased, and the motion detection signal KT (FIG. 2 I) finally obtained includes a moving image portion (38 ), and an unnecessary high-level "1" signal is not generated in the still image portion (39).

In this way, the inverted signal Δ of the delayed two-frame difference signal Δ,
, to force the output signal at position Pl to 0 to the zero level "°0" corresponding to the completely still image part.
For example, when an object moves from position PI to position P2 (
At the current time t, an object exists at position P2. ), position P2
This is because the position PI is determined to be a moving image part and the position PI is determined to be a still image part.

On the other hand, if only the 2-frame difference signal Δ4 (Fig. 2D) is used, not only the position Pt but also the position Pl will be determined to be a moving image part. The position Pl is determined to be a still image portion. In this sense, the delayed two-frame difference signal Δ can also be considered as a prohibition signal for the two-frame difference signal Δ4.

As described above, according to this example, by using the 2-frame difference signal Δ4 and the delayed 2-frame difference signal Δ, it is possible to accurately detect a moving image part, and especially when the input signal is 4M
There is an advantage that even if a high-frequency signal exceeding Hz is used, a still image portion will not be mistakenly determined to be a moving image portion. Therefore, it is possible to reduce blur in images reproduced by subsequent processing circuits.

Further, according to this example, since a sensitivity setting device (31) and a level setting device (34) are provided, the detection sensitivity of the moving image portion can be adjusted in various ways, and a still image portion is erroneously determined to be a moving image portion. There is an advantage in that the kind of sensitivity that suppresses this can also be adjusted in various ways.

Although the above-mentioned embodiment applies the present invention to the case where n=2, the present invention is not limited to this, and - n frames (n=2+ 3, 41 51 . . . ) in the film are applied. It is obvious that the present invention can be applied to a motion detection circuit for a television signal that uses the image signals of 1 to 2 to construct a complete image for one screen.

As described above, the present invention is not limited to the above-described embodiments, and it goes without saying that various configurations may be adopted without departing from the gist of the present invention.

〔Effect of the invention〕

According to the present invention, since only the difference signal between n frames is used, there is an advantage that a still image portion is not mistakenly determined to be a moving image portion not only for a low frequency signal but also for a high frequency signal.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram showing a motion detection circuit according to an embodiment of the present invention, FIG. 2 is a signal waveform diagram of each part to explain the operation of the example in FIG. 1, and FIG. 3 is a line diagram to explain the MOSR transmission method. 4 are block diagrams showing a conventional MtlSεt coder, and FIGS. 5 and 6 are signal waveform diagrams of various parts for explaining the operation of the example shown in FIG. 4, respectively. (28) is a 2-frame memory, (29) is a subtraction circuit, (
30) is the absolute value circuit, (31) is the sensitivity setting device, (33)
is a digital comparator, (35) is a 2 frame memory, (3
6) is an inverter circuit.

Claims (1)

[Claims] A motion detection circuit for a television signal configured to construct one screen image using image signals of n frames (n is an integer of 2 or more), comprising: a first memory that provides a delay of n frame periods; a difference means for determining the difference between the output signal of the first memory and the input television signal; and a difference means that provides a delay of n frame periods to the output signal of the difference means. and a second memory for giving a still image, and when the output signal of the second memory satisfies a predetermined condition, the output signal of the difference means is treated as a signal corresponding to a still image and processed. A motion detection circuit characterized by: