JPH01231487A - Moving signal processing circuit - Google Patents

Abstract

PURPOSE:To prevent mis-detection of a still picture as a moving picture due to the pattern of a picture by applying time spatial filter processing to a moving detection signal. CONSTITUTION:A moving detection signal detected based on an inter-frame difference signal of a digitized television signal is inputted and a time spatial filter 2 uses the both to apply filter processing in a direction of time and/or space. Moreover, when an output signal of the time spatial filter 2 is larger than a preset value, a control means 4 controls the signal so that a switch means 3 outputs an output signal of the time spatial filter means 2 and when the output signal is smaller than the preset value, the means 4 controls the signal so that the moving detection signal is outputted. Thus, when the output signal of the time spatial filter means 2 is larger than a value, the circuit is of feedback type configuration and the information of movement of the picture element over a wide range is referenced to detect the presence of the movement in the picture element on the screen. Thus, even to a fast movement of a picture, the movement is detected accurately.

Description

Detailed Description of the Invention [Industrial Application Field] The present invention relates to motion detection, which is detected based on an inter-frame difference signal of a digitized television signal and indicates the movement of an image obtained by the television signal. The present invention relates to a motion signal processing circuit that performs signal processing on a signal to obtain a motion control signal for controlling a motion adaptive signal processing circuit.

[Conventional technology]

Conventionally, there is a technique in which the increment scan performed in the NTSC system is converted into progressive scanning by interpolating scanning lines on the receiver side and displayed. In this way, if you use field memory to create an interpolated scanning line signal using the scanning line signal from one field before, and convert it to sequential scanning, you can eliminate line frizz that occurs at the edge when displaying a horizontal striped pattern. power can be removed. However, although this inter-field interpolation has a great effect on still images, it causes significant deterioration such as comb-like double images in the case of moving images.

Therefore, in the example shown in Japanese Patent Application Laid-Open No. 58-205377, the motion of the image is detected based on the difference signal between one frame, and if the motion of the image is small, interpolation is performed between fields to detect the motion of the image. When it is large, an interpolated scanning line signal is created using the transmitted scanning line signal within the field, and motion-adaptive processing is performed, such as performing interpolation between fields. However, when detecting the motion of an image based on the frame-to-frame difference signal in this way, there is a problem that the motion cannot be detected accurately if the motion of the image is fast, as described below. was there.

That is, for example, as shown in FIG.
When moving from top to bottom, the movement of the image J of an object is expressed using interlaced scanning lines, with the vertical axis being the vertical direction of the screen and the horizontal axis being the time axis.
It will look like Figure 7.

In FIG. 7, circles indicate the cross section of scanning line 2. In FIG.

When creating the interpolation scanning line indicated by the triangular mark in the M-th field, it is originally necessary to perform interpolation between the inner and outer fields since it is a moving image, but as is clear from FIG. (M-1) Scanning line I in field
! , M-1 and scan lines in the (M+1)th field!
. , 1, the difference becomes zero since the information on the defining scanning line is the same, and therefore, no movement in the position of the triangle mark can be detected from such an inter-frame difference signal. Therefore, interfield interpolation, which is still image processing, is performed on a moving image, resulting in deterioration such as the above-mentioned double image.

Further, for a fine pattern, even relatively slow movement may similarly result in erroneous motion detection.

Therefore, in order to solve this problem, in the previously proposed example described above, low-pass filter processing is performed on the motion of the image, that is, the amount of motion, detected based on the difference signal between one frame. Compatible.

For example, when an image J of an object moves from left to right on the screen, the movement of the image J of the object is expressed by taking the vertical axis as the time axis and the horizontal axis as the horizontal direction of the screen, as shown in Figure 8 (a). ). Then, as the inter-frame difference signal, the (M-1)th
The amount of motion detected by taking the difference between the field and the (M+1)th field is as shown in FIG. 8(b). but,
In this case, even though there is always movement, there is a range where the amount of movement is zero, resulting in false detection. Therefore, processing for correction, that is, low-pass filter processing, is performed on this amount of motion, and the amount of motion after correction is obtained as shown in FIG. 8(c).

However, as described below, with this method, similar patterns of movement may occur depending on the image, and because video processing is performed on still images, there is a possibility that the resolution may decrease or flicker may occur. was there.

That is, for example, as shown in FIG. 9(a), an image S of an object is stationary, and two frame images W, , W.

If only the object slowly moves from left to right on the screen, the detected amount of movement will be the same as shown in FIG. 9(b), as shown in FIG. As shown in c), it will be similar to Fig. 8(c), so between the two frames,
The image S of the object is processed as a moving image even though it is a still image (i.e., inside the field).
You end up doing this.

In addition, in the above-mentioned general proposal example, whether or not false detection can be prevented up to the speed of movement is preset by the circuit constant. does not occur.

[Problem that the invention seeks to solve]

As mentioned above, in the conventional technology, in order to accurately detect fast movements in images, low-pass filter processing is performed on the amount of movement detected based on the inter-frame difference signal, that is, the motion detection signal. However, depending on the image design, still images may be mistakenly detected as moving images. Further, although a certain level of fast movement can be detected accurately, there is a problem in that erroneous detection occurs for faster movement.

Furthermore, in the above-mentioned conventional technology, due to the influence of noise and the like, there may be cases where motion is erroneously detected even though there is no motion.

The present invention has been made in view of the problems of the prior art described above, and therefore, its purpose is to accurately detect the movement of an image regardless of the image pattern, etc., and to detect fast movement of the image. It is an object of the present invention to provide a motion signal processing circuit that can accurately detect motion signals and that is less likely to cause erroneous detection due to the influence of noise or the like.

[Means to solve the problem]

In order to achieve the above object, the present invention inputs a motion detection signal detected based on an interframe difference signal of a digitized television signal and a certain first signal,
A spatio-temporal filter means that performs filter processing in the temporal direction and/or spatial direction using both of them, and a spatio-temporal filter means that inputs the motion detection signal and the output signal of the spatio-temporal filter means, switches between them, and performs filter processing in the temporal and/or spatial direction. A switching means inputs the first signal to the spatio-temporal filter means, and which signal is outputted by the switching means depending on whether the value of the output signal of the spatio-temporal filter means is larger or smaller than a certain value. and a control means for controlling.

[Effect]

The spatio-temporal filter inputs the motion detection signal and the output signal of the switching means (i.e., the first signal), and uses both of them to detect the motion in the temporal direction and/or the spatial direction (vertical and horizontal directions of the screen). Perform filter processing on

The switching means inputs the motion detection signal and the output signal of the spatio-temporal filter means, and outputs either one under switching control by the control means.

The control means controls the switching means to output the output signal of the spatio-temporal filter means when the value of the output signal of the spatio-temporal filter means is larger than a certain value set in advance. , the switching means is controlled to output the motion detection signal.

Therefore, when the value of the output signal of the spatio-temporal filter means is larger than the certain value, the feedback type configuration is adopted, and therefore it is necessary to detect the presence or absence of movement at a certain pixel on the screen over a wide range. It is possible to refer to information on the movement of pixels (spatiotemporally surrounding pixels, i.e., up and down, left and right, and past pixels), and therefore it is possible to accurately detect movement even when the image moves quickly. Can be done.

Furthermore, when the value of the output signal of the spatio-temporal filter means is smaller than the certain value, a feed-forward type structure is adopted, in which the information on the movement once detected persists over several fields, or the information on the surrounding It does not spread to other pixels. Therefore, even if motion is mistakenly detected even though there is no motion due to the influence of noise, that information is immediately attenuated and does not spread to surrounding pixels or persist. Erroneous detections due to influence can be minimized.

Note that the effect of accurately detecting fast motion in an image using the above-mentioned spatio-temporal filter has been filed by the present inventors in Special Training No. 60-250912.

〔Example〕

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

In FIG. 1, 1 is an input terminal, 2 is a spatiotemporal filter circuit, 3 is a switching circuit, 4 is a control circuit, 5 is a coefficient circuit, 6 is an output terminal, and 7 is a field memory.

8 is a line memory, 9 is a maximum value selection circuit, and 10 is a horizontal filter circuit. In addition, the following explanation is for the sake of simplicity.
TSC television signal (525 scanning lines, 2
: 1 increment scan). In this case, the field memory 7 may have a 262H capacity (IH is one horizontal scanning period), and the line memory 8 may have an LH capacity.

Now, a motion detection signal (this signal is a digital signal), which is detected based on the interframe difference signal of the digitized television signal and indicates the movement of the image obtained by the television signal, is input from the input terminal 1. It is manually supplied to the spatio-temporal filter circuit 2 and the switching circuit 3 respectively. The switching circuit 3 switches between this motion detection signal and a signal obtained by multiplying the output signal of the spatio-temporal filter circuit 2 by α (0<α, <1) by the coefficient circuit 5, and supplies the signal to the spatio-temporal filter circuit 2.

In the spatiotemporal filter circuit 2, the output signal from the switching circuit 3 is delayed by the field memory 7 and the line memory 8, and a signal delayed by 262H and a signal delayed by 263H are obtained, respectively. Then, by the maximum value selection circuit 9,
Among the two delayed signals and the motion detection signal from the input terminal 1, the maximum signal value (that is, the level value of the digital signal) is determined and output. Then, the horizontal filter circuit 10 performs filter processing in the horizontal direction of the screen, and outputs from the output terminal 6.

The signal output from the output terminal 6 is used as a motion control signal for controlling a motion adaptive signal processing circuit (not shown).

Furthermore, when the value of the output signal of the spatio-temporal filter circuit 2 is smaller than a preset value, the control circuit 4 selects the lower side of the switch of the switching circuit 3 and outputs the motion detection signal from the input terminal l. The switching circuit 3 is controlled to output the signal from the coefficient circuit 5, and when the value is large, the switch selects the upper side and outputs the output signal from the coefficient circuit 5.

The above operation will be explained in more detail using FIG. 2.

In Figure 2, as in Figure 7, when the object image J moves from the top to the bottom of the screen, interlaced scanning is performed with the vertical axis in the vertical direction of the screen and the horizontal axis in the time axis. FIG. 3 is an explanatory diagram showing the movement of an image J of an object using scanning lines.

In Figure 2, the circles are scanning lines! A cross section is shown.

Now, in FIG. 1, it is assumed that the switch of the switching circuit 3 is selected upward by the control circuit 4, and that the input terminal 1 is connected to the triangular mark (
It is assumed that a motion detection signal at the position of n+ (interpolated scanning line) is input. That is, the motion detection signal at the position of the triangular mark n is obtained by taking the difference between the scanning line 2I in the (M+1)th field and the scanning line 24 in the (M-1)th field, as an interframe difference signal. It is something that can be done.

Now, let's consider what kind of signals are being obtained from the field memory 7 and line memory 8 when the motion detection signal at the position of the triangle mark n1 is being input from the input terminal 1. View.

First, the first thing to think about is that since the field memory 7 and line memory 8 are delayed by 262H and 263H, the motion detection signal at the input triangular mark n is input 262H or 263H before. This is the motion detection signal.

That is, the motion detection signal before 262H is the motion detection signal at the position of the triangle mark n3 in FIG. 2 (obtained by taking the difference between scanning line 2. and scanning line I!, 6),
The previous motion detection signal is the motion detection signal at the position of the triangle mark n2 (obtained by taking the difference between scanning line !2 and scanning line 25).
It is. However, these two motion detection signals are not output as they are from the field memory 7 or line memory 8, but instead pass through the coefficient circuit 5.
Each motion detection signal is multiplied by α and output.

The next possibility is (262H+262H), (262H+263H) or (263H+263H) before the motion detection signal at the input triangular mark nl.
) is the previously input motion detection signal. That is, (262
The motion detection signal before (H+262H) is the motion detection signal at the position of the triangle mark n6 in FIG.
+263H) previous motion detection signal is a motion detection signal at the position of triangle mark n, and (263)1+263)I) previous motion detection signal is a motion detection signal at the position of triangle mark n4. Furthermore, since the signal passes through the coefficient circuit 5 twice, each motion detection signal is outputted from the field memory 7 or line memory 8 in the form multiplied by α×α.

Another possibility is that the motion detection signal at the position of each triangular mark in the (M=3)th field shown in FIG.
This is a signal multiplied by three.

In this way, from the field memory 7 and the line memory 8, signals α, αZ, which are about one field before, about two fields before, etc., are sent from the input motion detection signal at the position of the triangular mark n1. It is conceivable that the output is weighted as follows.

Next, the maximum value selection circuit 9 selects the motion detection signal at the position of the triangle mark n from the input terminal 1 and the output signals from the field memory 7 and line memory 8, which have the largest signal value. will be output. That is, in FIG. 2, the triangle mark n.

The motion detection signal at the position of is zero (because if you take the difference between scanning line 1. and scanning line 14, the information on the re-scanning line is the same, so the difference is zero), but at the position of triangular mark nff1 + triangular mark n3 Each position motion detection signal has a certain value (if we take the difference between scanning line 12 and scanning line 2, or the difference between scanning line 1 and scanning line 16, the information on each re-scanning line is different. (the difference has a certain value), so the signal multiplied by α also has a certain value.

Also, triangle mark n4. Since the motion detection signals at the positions of the triangular mark ns+triangular mark n6 are each zero, the signal α2 times the signal is also zero. Therefore, the maximum value selection circuit 9 outputs a motion detection signal multiplied by α at the position of the triangular mark n2 or the triangular mark n3.

That is, conventionally, as described in FIG. 7, the position of the triangular mark in the Mth field (triangular mark n in FIG. 2).

(corresponding to the position j), which was erroneously detected as "no movement", but in this embodiment, it is accurately detected as "movement j".

As described above, in this embodiment, when the switch of the switching circuit 3 is set to the upper side, the presence or absence of movement at a certain pixel on the screen is detected by detecting the presence or absence of movement at that pixel (in the above example, the triangular mark n). It is determined by also referring to information on the movement of pixels in the vertical direction (in the vertical direction of the screen) and in the past.

Next, the output signal from the maximum value selection circuit 9 is filtered in the horizontal direction of the screen in the horizontal filter circuit 10 as described above. As a result, detection of the presence or absence of movement at a certain pixel on the screen is determined by referring not only to the above-mentioned upper, lower, and past pixels, but also to information about the movement of pixels to the left and right (horizontal direction of the screen). will be done.

Here, when detecting the presence or absence of movement in a certain pixel on the screen, the range of pixels that can refer to movement information is the range determined by the coefficient α (0 < α × 1) in the coefficient circuit 5. It is. That is, the closer α is to 1, the wider the referenceable pixel range becomes, and the closer α is to 0, the narrower it becomes.

In addition, in the above explanation, it was explained that the coefficient circuit 5 multiplies the human input signal of the coefficient circuit 5 by α (0 < α * 1) and outputs it, but since both the input signal and the output signal are digital signals,
Even if the input signal is accurately multiplied by α, the output signal will be approximated to an integer. Therefore, since the output signal is output in an approximated form, subtraction processing may be performed as an approximate calculation instead of multiplying by α. For example, when α=0.9, instead of multiplying by 0.9, subtraction processing according to the formula of output signal = input signal - 1 (however, when input signal ≧ 1) is performed, As a result, substantially similar output signals will be obtained.

Next, in FIG. 1, the switch of the switching circuit 3 is selected to the lower side by the control circuit 4, and the input terminal 1
From here is the triangle mark n shown in Figure 2.

A case will be explained in which a motion detection signal at the position is input.

In this case, field memory 7 and line memory 8
262 from the motion detection signal at the position of the triangle mark n1.
The motion detection signal input before H and before 263H, that is,
Only the motion detection signal at the position of the triangular mark n and the motion detection signal at the position of the triangular mark n2 are output.

Therefore, the maximum value selection circuit 9 outputs the maximum signal value among the motion detection signals at each position of the triangular mark nl + the triangular mark n2 + the triangular mark n, and the horizontal filter circuit 10
It reaches the output terminal 6 via.

That is, in this embodiment, when the switch of the switching circuit 3 is selected to the lower side, the detection of the presence or absence of movement at a certain pixel on the screen is performed at that pixel (in the above example, the pixel at the position of the triangular mark n1). ), the decision is made by referring only to information about the movement of adjacent vertical, horizontal, and past pixels.

Next, how the switch of the switching circuit 3 is switched by the control circuit 4 in accordance with the temporal change in the value of the output signal of the spatio-temporal filter circuit 2 will be explained using FIG.

FIG. 3 is a characteristic diagram showing temporal changes in the value of the output signal of the spatio-temporal filter circuit in FIG. 1 when focusing on a certain pixel on the screen.

In FIG. 3, the vertical axis indicates the value of the output signal of the spatio-temporal filter circuit 2 for a certain pixel on the screen, and the horizontal axis indicates time. Note that the time shown on the horizontal axis is 1/
It is divided into 60 seconds, the first division is the first field, the second division is the second field,
The third delimiter is the third field.

In FIG. 3, starting from the left, the signal values at input terminal 1 are 2.4°, 6.8, 10, .
12 shows temporal changes in the value of the output signal of the spatio-temporal filter circuit 2 when a signal of 12 is input in the first field. It is assumed that the input motion detection signal of a certain pixel has a signal value of O before the first field and after the second field.

Further, in FIG. 3, the value set in the control circuit 4 is 5, and the operation in the coefficient circuit 5 is shown as performing a subtraction process according to the formula: output signal=input signal-2.

Therefore, for example, if a signal with a signal value of 12 is input to input terminal 1 as a motion detection signal for a certain pixel in the first field, and signals with a signal value of O in the second and subsequent fields are input, the motion detection signal for that pixel is The value of the output signal of the spatio-temporal filter circuit 2 changes over time as shown on the rightmost side of FIG.

That is, in the first field, the spatio-temporal filter circuit 2 outputs the motion detection signal from the input terminal l, that is, a signal whose signal value is 12. At this time, since the value of this output signal is larger than the value of 5 that has been set in advance in the control circuit 4, the control circuit 4 switches the switch of the switching circuit 3 to the upper side. From the second field onward, the value of the motion detection signal from input terminal 1 is 0, so
In place of the motion detection signal from the input terminal 1, the coefficient circuit 5
．． Field switching circuit 3. A signal via field memory 7 (or line memory 8 as well) is output. However, since the value of this signal is subtracted by 2 by the coefficient circuit 5, the value of the signal decreases by 2 each time it passes through the coefficient circuit 5, that is, for each field.

Then, at the time of the fifth field, the spatio-temporal filter circuit 2
The value of the output signal is 4, which is smaller than the value of 5 that has been set in advance in the control circuit 4, so the control circuit 4 switches the switch of the switching circuit 3 to the lower side. As a result, in the fifth field, the spatio-temporal filter circuit 2 outputs a signal whose value stored in the field memory 7 is 4, but at the same time, the value from the input terminal 1 is output to the field memory 7. The signal (motion detection signal) will be written. Therefore, in the 6th field,
The value of the output signal of the spatio-temporal filter circuit 2 becomes 0.

For example, if a signal whose signal value is 4 is input to the input terminal 1 as a motion detection signal of a certain pixel in the first field, and a signal whose value is 0 is input to the second field, The value of the output signal of the spatio-temporal filter circuit 2 for that pixel changes over time as shown in the second row from the left in FIG.

That is, in the first field, the space-time filter circuit 2 outputs the motion detection signal from the input terminal l, that is, a signal whose signal value is 4, and the switch of the switching circuit 3 selects the lower side. Therefore, a signal having a value of 4 (motion detection signal) from the input terminal 1 is written into the field memory 7. However, in the second field,
Since the value of the motion detection signal from the input terminal 1 is 0, the signal written in the field memory 7 with a value of 4 is output from the spatiotemporal filter circuit 2, and the signal from the input terminal 1 is output from the field memory 7. A signal (motion detection signal) with a value of O will be written.

Therefore, in the third field, the value of the output signal of the spatio-temporal filter circuit 2 is 0.

Now, Fig. 3 is a characteristic diagram showing the temporal change in the value of the output signal of the spatiotemporal filter when focusing on a certain pixel on the screen. If we take the vertical or horizontal direction of the screen instead of the time, the diagram shows the spatiotemporal filter for a certain pixel and each pixel adjacent to that pixel along the vertical or horizontal direction of the screen. It will represent the value of the output signal.

However, in this case, the values at each pixel are not at the same time.

As explained above, in this embodiment, instead of performing low-pass filter processing like the conventional one, filter processing is performed by the spatio-temporal filter circuit 2.
, there is no possibility that a still image will be mistakenly detected as a moving image due to the image pattern.

Further, when the value of the output signal of the spatio-temporal filter circuit 2 is larger than the value set in advance in the control circuit 4, the output signal from the coefficient circuit 5 is output from the switching circuit 3, and the entire circuit is Since it has a feedback type configuration and can refer to information on the movement of a wide range of pixels, it is possible to accurately detect movement even when the image moves quickly.

Further, when the value of the output signal of the spatio-temporal filter circuit 2 is smaller than the value set in advance in the control circuit 4, the motion detection signal from the input terminal 1 is output from the switching circuit 3, and the entire circuit is has a feedforward type configuration, and the range of pixels that can be referenced is limited to adjacent pixels, so even if motion is mistakenly detected even though there is no motion due to the influence of noise, etc.
This information can be prevented from spreading to surrounding pixels.

Therefore, false detections due to the influence of noise etc. can be minimized.

Note that the values set in advance in the control circuit 4 are as follows:
Try to actually evaluate the image and choose the optimal value.

Now, in the embodiment shown in FIG. 1 described above, the line memory 8 is connected in cascade after the field memory 7 when obtaining the 262H delayed signal and the 263H delayed signal, but the present invention is not limited to this. do not have. For example, as shown in FIG. 4, the field memory 7 may be arranged after the line memory 8.

FIG. 4 is a block diagram showing a second embodiment of the present invention.

In FIG. 4, 11 and 12 are maximum value selection circuits, and the others are the same as in FIG. 1.

In this embodiment, the maximum value selection circuit 11 determines in advance the larger signal value of the motion detection signals of two pixels separated by IH, that is, vertically adjacent pixels, and outputs the signal from the maximum value selection circuit 11. is supplied to the field memory 7 and 26
By delaying by 2H, the output signal from the switching circuit 3 is outputted from the field memory 7 by 262H and 263H.
Among the delayed signals, the one with the larger signal value is obtained.

The maximum value selection circuit 12 selects the motion detection signal from the input terminal l and the output signal from the field memory 7, whichever has a larger signal value. Therefore, maximum value selection circuit 1
2, the maximum signal value is obtained from the motion detection signal from the input terminal l and the signal obtained by delaying the output signal from the switching circuit 3 by 262H and 263H, which is shown in Figure 1. This is equivalent to the output of the maximum value selection circuit 9 in the embodiment.

The operations of other parts in this embodiment are equivalent to those in the embodiment shown in FIG. 1, and therefore their explanation will be omitted.

Next, FIG. 5 is a block diagram showing a third embodiment of the present invention.

In FIG. 5, numeral 13 is a switching control circuit, and the others are the same as in FIG. 1.

The switching control circuit 13 in this embodiment is configured by integrating the switching circuit 3, the control circuit 4, and the coefficient circuit 5 in the embodiment shown in FIG. The output signal of the spatial filter circuit 2 is used as an address input, and a ROM (read/write) outputs data.
Only Mesori).

That is, the switching control circuit 13 inputs the output signal of the spatio-temporal filter circuit 2, and when the value of the output signal of the spatio-temporal filter circuit 2 is large, the output signal of the spatio-temporal filter circuit 2 is multiplied by α and outputted. Further, when the value of the output signal of the spatio-temporal filter circuit 2 is small, it operates to output a motion detection signal from the input terminal 1.

According to this embodiment, the configuration can be simplified by integrating the switching circuit 3, control circuit 4, and coefficient circuit 5 in FIG. 1 into the switching control circuit 13. In addition, by using a ROM as the switching control circuit 13, it is possible to output an arbitrary value, and by setting an appropriate value based on the results obtained from actual image quality evaluation, the motion detection performance can be adjusted. It is possible to further improve the performance.

(Effects of the Invention) In the present invention, instead of performing low-pass filter processing as in the past, spatiotemporal filter processing is performed on the motion detection signal detected based on the interframe difference signal of the television signal. Therefore, unlike in the past, a still image is not mistakenly detected as a moving image due to the image pattern.

Further, according to the present invention, when detecting the presence or absence of movement at a certain pixel on the screen, if the output signal of the spatiotemporal filter means is large, the spatiotemporal filter processing is a feedback type processing. Information on the movement of a wide range of pixels (spatiotemporally surrounding pixels, i.e., up/down, left/right, and past pixels) can be referenced, making it possible to accurately detect movement even in fast moving images. can do. Also,
When the output signal of the spatio-temporal filter means is small, the spatio-temporal filter processing is a feed-forward type processing, and even if motion is erroneously detected even though there is no motion due to the influence of noise etc. Since the information is immediately attenuated, the information does not spread to surrounding pixels or persist, making it possible to minimize false detections due to the influence of noise and the like.

Therefore, there is an effect of improving image motion detection performance.

[Brief explanation of the drawing]

Figure 1 is a block diagram showing the *1 embodiment of the present invention;
The figure is an explanatory diagram for explaining the operation of the spatio-temporal filter circuit in Fig. 1, and Fig. 3 shows the value of the output signal of the spatio-temporal filter circuit in Fig. 1 when focusing on a certain pixel on the screen. A characteristic diagram showing temporal changes, FIG. 4 is a block diagram showing the second embodiment of the present invention, FIG. 5 is a block diagram showing the third embodiment of the present invention, and FIG. FIG. 7 is an explanatory diagram showing the movement of the object image in FIG. 6 using scanning lines;
Figure 8 (a) is an explanatory diagram showing the image of an object moving from left to right on the screen, and Figure 8 (b) is the movement obtained by the movement of the image of the object in Figure 8 (a). Figure 8 (c) is an explanatory diagram showing the amount of movement after correction obtained by correcting the amount of movement in Figure 8 (b), and Figure 9 (a) is the left side of the screen. An explanatory diagram showing the image of a frame moving toward the right from the center and the image of a stationary object. Figure 9 (b) shows the amount of movement obtained by the movement of the image of the frame in Figure 9 (a). Explanatory diagram,
FIG. 9(c) is an explanatory diagram showing the corrected amount of motion obtained by correcting the amount of motion in FIG. 9(b). Explanation of symbols 1... Input terminal, 2... Space-time filter circuit, 3.
...Switching circuit, 4...Control circuit, 5-...Coefficient circuit,
6...Output terminal, 7...Field memory, 8...
- Line memory, 9, 11.12... Maximum value selection circuit, 10... Horizontal filter circuit, 13... Switching control circuit. Agent Patent Attorney Akio Namiki 1-Route Figure 8 (a) (b) (C) Figure 9 (Q'J (b) (C)

Claims (1)

[Claims] 1. A motion detection signal that is detected based on an interframe difference signal of a digitized television signal and indicates the motion of an image obtained by the television signal is subjected to signal processing to detect the motion. In a motion signal processing circuit that obtains a motion control signal for controlling an adaptive signal processing circuit, the motion detection signal and a certain first signal are input, and both are used to perform a motion control signal in a temporal direction and/or a spatial direction. spatio-temporal filter means for performing filter processing to obtain the motion control signal; and spatio-temporal filter means for inputting the motion detection signal and the motion control signal, switching between them and using one as the first signal. and a control means for controlling which signal is output by the switching means depending on whether the value of the motion control signal is larger or smaller than a certain value. A motion signal characterized in that the attenuation rate in the temporal direction and/or spatial direction of the motion control signal due to the filter processing of the spatio-temporal filter means can be changed according to the value of the motion control signal by control. processing circuit. 2. The motion signal processing circuit according to claim 1, wherein the switching means and the control means are integrated.