JPH06339124A - Motion detector - Google Patents

Abstract

PURPOSE:To reduce production of a missing spot without incurring generation of comb-line distortion even in a horizontal edge with a small luminance difference. CONSTITUTION:An absolute value ¦FD¦ of an inter-frame difference FD calculated by an inter-frame difference calculation circuit 211 and threshold levels TH, TL outputted from a threshold level generating circuit 212 are compared at a picture discrimination circuit 213. Thus, Whether a picture is an animation or a still picture is decided. Furthermore, the threshold levels TH, TL are controlled to be increased when a magnitude ¦V¦ of a motion vector detected by a motion vector detection section 20 is smaller than the threshold level TV. Thus, a picture is easily decided to be a still picture by the picture decision circuit 213.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion detecting device for determining whether an image is a moving image or a still image in a television standard conversion device, for example.

[0002]

2. Description of the Related Art Recently, due to the popularization of consumer video tape recorders and the enhancement of international communication networks, the exchange of television programs of different systems has increased. Along with this, the demand for devices that convert the television standard system has been increasing year by year.

One of the conversion items of the television standard system is conversion of the number of scanning lines. This conversion is basically achieved by thinning out one line or repeating one line for each predetermined number of lines.

However, in such a system, discontinuous distortion of an image occurs in a portion where the number of lines is increased or decreased.

In order to solve this problem, conversion of the number of scanning lines by a line interpolation method has been conventionally considered.

[0006] As this system, "Television Society Journal" Image Information Engineering and Broadcasting Technology Vol. 45 No. 12
1991 pp. 40-49 ", there are" intra-field line interpolation method "," inter-field line interpolation method "," inter-field / intra-field adaptive line interpolation method "and the like.

Here, the intra-field line interpolation method is, for example, a method of obtaining an interpolated signal by linear interpolation using two adjacent lines in the field. The inter-field line interpolation method is a method for obtaining an interpolated signal by linear interpolation using two adjacent lines between fields, for example. In addition, the inter-field / intra-field adaptive line interpolation method determines whether an image is a moving image or a still image,
This is a method of adaptively switching between the inter-field line interpolation method and the intra-field line interpolation method based on the determination result.

The intra-field line interpolation method can reduce the discontinuous distortion of the image, but has a problem that the vertical resolution is deteriorated.

On the other hand, in the inter-field line interpolation method, the deterioration of the vertical resolution can be reduced to 1/2 as compared with the intra-field line interpolation method. This is because the correlation between two adjacent lines between fields is greater than the correlation between two adjacent lines within a field.

This will be described with reference to FIG.

In the figure, A and B are video signals for two adjacent lines in the current field, C is a video signal in the previous field, and D is an interpolation signal.

In the intra-field line interpolation method, the interpolated signal D is obtained by weighted addition by multiplying the video signals A and B for two adjacent lines in the current field by the line interpolating coefficients α and (1-α). Desired. On the other hand, in the inter-field line interpolation method, the line interpolation coefficient (1-β), is added to the signals B and C of two adjacent lines of the current field and the previous field.
The interpolated signal D is obtained by the weighted addition multiplied by β.

Here, when comparing the spatial distance between the video signals A and B and the spatial distance between the video signals B and C, the latter is ½ of the former. Therefore, it is possible to reduce the deterioration of the vertical resolution by obtaining the interpolated signal D using the video signals B and C.

However, this effect holds only in the case of a still image, and in the case of a moving image, as shown in FIG. 3, the vertical edge of the moving object is distorted in a comb shape.

The inter-field / intra-field adaptive line interpolation method is made by paying attention to this point. In-field line interpolation is performed in the moving image area, and inter-field line interpolation is performed in the still image area. This reduces the occurrence of discontinuous distortion and deterioration of vertical resolution.

In the inter-field / intra-field adaptive line interpolation method, a digital filter is used as the inter-field line interpolation method and the intra-field line interpolation method instead of the method using two lines.
A method using more lines may be adopted.

By the way, in order to realize the above-mentioned inter-field / in-field adaptive line interpolation method, a motion detecting device for judging whether an image is a moving image or a still image is required.

This motion detecting device normally determines an inter-frame difference value of a video signal for each pixel and compares it with a predetermined threshold value to determine whether the image is a moving image or a still image. Has become. That is, if the inter-frame difference value is smaller than the motion detection threshold value, it is determined to be a still image,
If it is large, it is determined to be a moving image.

In such a configuration, the image quality of the interpolated image is greatly influenced by the motion detection accuracy of the motion detection device. Especially when the video is mistakenly judged as a still image,
As described above, the comb distortion occurs. Therefore, in the motion detection device, it is desired to prevent such an erroneous determination.

In order to meet this demand, conventionally, there has been considered a method of setting the motion detection threshold value to a value that makes it easy to determine that it is a moving image.

According to such a method, it is difficult to determine that a moving image is a still image, so that it is possible to reduce the occurrence of comb distortion.

However, in such a method, there is a new problem that the worm-eating phenomenon shown in FIG. 4 may occur at the horizontal edge portion having a small vertical correlation.

That is, consider a case where the amount of motion changes within one line or less at the horizontal edge portion having a small vertical correlation. In this case, unless the motion detection threshold value is set to a value that makes it easy to determine that it is a moving image, the inter-frame difference value rarely exceeds the motion detection threshold value, and the determination result of the motion detection device is a still image.

However, if the motion detection threshold value is set to a value that makes it easy to determine that it is a moving image, the inter-frame difference value may or may not exceed the motion detection threshold value. As a result, the determination result of a moving image and the determination result of a still image are mixedly obtained.

When such a state occurs, inter-field line interpolation and inter-field line interpolation are performed in a mixed manner, so that the above-mentioned worming phenomenon occurs due to the difference between the corrected images of both.

In order to solve this problem, conventionally, a method of detecting an edge in a frame and suppressing the interframe difference value based on the edge detection result has been considered.

According to such a method, even if the amount of motion changes in one line or less at the horizontal edge portion having a small vertical correlation, the inter-frame difference value does not easily exceed the motion detection threshold value. It is possible to reduce the occurrence of the insect-eating phenomenon.

[0028]

As described above, in the conventional motion detecting device, in order to reduce the occurrence of the worm-eating phenomenon in the horizontal edge portion having a small vertical correlation, the edge in the frame is detected. The interframe difference value is suppressed based on the detection output.

However, with such a structure, the worming phenomenon can be reduced at the horizontal edge portion having a large luminance difference, but cannot be reduced at the horizontal edge portion having a small luminance difference (eg, stairs). There was a problem. This is because the edge cannot be detected in such a portion.

That is, the edge is detected by comparing the video signal with a predetermined threshold value. In this case, in the horizontal edge portion having a large brightness difference, the horizontal edge can be reliably detected even if the edge detection threshold value is set to an appropriate size. On the other hand, in the edge portion where the brightness difference is small, the horizontal edge cannot be detected unless the threshold value for edge detection is made considerably small.

However, if the threshold value for edge detection is made too small, the effect of suppressing the inter-frame difference value will work even in portions other than horizontal edges. As a result, it is erroneously determined as a moving image to be a still image, and the above-described distortion occurs.

Therefore, the edge detection threshold cannot be made too small. As a result, conventionally, it was not possible to reduce the worm-eating phenomenon at the horizontal edge portion where the brightness difference is small.

The present invention has been made in order to cope with the above situation, and it is possible to reduce the occurrence of worm-eating phenomenon even at a horizontal edge portion having a small luminance difference without causing comb distortion. The purpose is to provide a device.

[0034]

In order to achieve the above object, the present invention compares an amplitude difference of video signals separated by N (N is an integer of 1 or more) fields with a predetermined threshold value to obtain an image. In a motion detection device that determines whether a moving image or a still image, if the size of the motion vector indicating the magnitude and direction of the motion of the moving image is small, the size of the motion vector is set so that the image is easily determined as a still image. The image determination operation is controlled based on the above.

[0035]

In the above structure, the image determining operation is controlled so that the image is more likely to be determined as a still image when the magnitude of the motion vector becomes smaller.

Here, the magnitude of the motion vector becomes small in the horizontal edge portion having a small vertical correlation and in the area where the motion amount changes within one line or less. Therefore, in such an area, by the control as described above,
The image is more likely to be judged as a still image, and the occurrence of the worm eating phenomenon is reduced.

Moreover, the small motion vector means that the image is close to a still image. Therefore,
Even if the above-described control is performed, there is almost no malfunction that a moving image is determined to be a still image. As a result, even if the difference in brightness at the horizontal edge portion is small, it is possible to reduce the occurrence of the worming phenomenon without causing comb distortion.

[0038]

Embodiments of the present invention will now be described in detail with reference to the drawings.

First, a first embodiment of the present invention will be described. In the following description, the case where the present invention is applied to a motion detection device for switching between inter-field line interpolation and intra-field line interpolation will be described as a representative.

FIG. 1 is a block diagram showing the overall configuration of a television standard system conversion apparatus which employs inter-field line interpolation / inter-field line interpolation as a line interpolation system.

Referring to FIG. 1, first, the overall structure and operation of the television standard conversion device will be described.

Video signal S1 input from the input terminal 11
Is supplied to the motion correction field interpolation unit 12 and is also supplied to the field memory 13. The video signal S2 supplied to the field memory 13 is delayed by one field and then supplied to the motion compensation field interpolation unit 12 as the video signal S2.

The video signals S1 and S2 supplied to the motion correction field interpolation unit 12 are motion-corrected using a motion vector V detected by a motion vector detection unit, which will be described later, and then interpolated in the field.

The video signal S3 obtained by this field interpolation is supplied to the interframe interpolation section 14 and delayed by one frame by the two field memories 141 and 142. The video signal S5 and the video signal S3 obtained by this delay are multiplied by the multiplication circuits 143, 144 and the addition circuit 14 respectively.
5, the weighted addition is performed according to the inter-frame interpolation coefficient of 1/2. As a result, the video signal S6 interpolated between frames
Is obtained.

The video signal S6 is supplied to the inter-field line interpolating section 15 together with the video signal S4 obtained by delaying the video signal S3 by one field by the field memory 141.

The video signals S6 and S4 supplied to the inter-field line interpolating unit 15 are inter-field line interpolating coefficients (1-β) and β (0≤β≤ by the multiplying circuits 151 and 152 and the adding circuit 153. Weighted addition is performed according to 1). Thereby, the video signal S7 interpolated between the fields is obtained.

The video signal S4 is supplied to the in-field line interpolating section 16 and delayed by one line by the line memory 161. The video signal S8 obtained by this delay and the video signal S4 are multiplied by the multiplication circuits 162, 16
3 and the addition circuit 164 perform weighted addition according to the inter-field line interpolation coefficient (1-α), α (0 ≦ α ≦ 1). As a result, the video signal S9 interpolated in the field line is obtained.

The video signals S7 and S9 are sent to the switching unit 17
To the multiplying circuits 171, 172 and the adding circuit 173.
Thus, the weighted addition is performed according to the inter-field / in-field line interpolation coefficient γ (0 ≦ γ ≦ 1), (1-γ). As a result, the video signal S10 inter-field / in-field adaptive line interpolated is obtained. The video signal S10 is output from the output terminal 18.

The video signal S1 input from the input terminal 11 is further supplied to the motion vector detecting section 20 and the motion detecting section 21 which is a feature of the present invention. The video signal S2 output from the field memory 13 further includes the field memory 19 and the motion vector detecting section 20.
Is supplied to. The video signal S2 supplied to the field memory 19 is delayed by one field, and then the video signal S1 is delayed.
1 is supplied to the motion detection unit 21.

The motion vector detecting section 20 detects the video signal S
1 and S2 are used to detect the motion vector V indicating the magnitude and direction of the motion of the moving image. In this detection, the screen for one field is m pixels x n lines (m and n are integers of 1 or more)
It is divided into blocks having the size of and is made for each block. Further, as a detection method in this case, a method using an iterative gradient calculation or the like is used.

The detected motion vector V is supplied to the motion correction field interpolation unit 12 and the motion detection unit 21 which is a feature of the present invention. In the motion detector 21, the video signal S1
Then, by using the video signal S11 delayed by one frame and the motion vector V, it is determined whether the input video is a moving image or a still image.

The result of this determination is output as the inter-field / intra-field adaptive line interpolation coefficient γ, (1-γ). The inter-field / intra-field adaptive line interpolation coefficients γ, (1-γ) are delayed by one field by the field memory 22 for time alignment with the video signal S10 to be interpolated, and then each of the multiplication circuits 171. , 1
72.

The above is the overall configuration and operation of the television standard conversion apparatus. Next, the configuration of the motion detector 21 which is a feature of the present invention will be described.

The motion detecting section 21 comprises an inter-frame difference value calculating circuit 211, a threshold value generating circuit 212, an image judging circuit 213, and a threshold value controlling circuit 214.

The inter-frame difference value calculation circuit 211, for each pixel, the difference in amplitude between the video signal S1 input from the input terminal 11 and the video signal S11 one frame before output from the field memory 19, that is, An interframe difference value FD is obtained.

The threshold generation circuit 212 receives the input video signal S1.
Thresholds T for determining whether the image of the image is a moving image or a still image
L and TH are generated. Here, this motion detection threshold value T
The magnitude relationship between L and TH is set to TL <TH.

The image determination circuit 213 determines, for each pixel, whether the image is a moving image or a still image by comparing the absolute value | FD | of the inter-frame difference value FD with the thresholds TL and TH. The result of this determination is output as the inter-field / intra-field adaptive line interpolation coefficient γ, (1-γ).

The threshold control circuit 214 compares the magnitude | V | of the motion vector V detected by the motion vector detecting section 20 with a predetermined threshold TV, and based on the comparison result, the motion detection threshold TL, Control TH.

That is, the threshold control circuit 214 has | V |
Is larger than TV, TL and TH are made smaller so that it can be easily determined as a moving image, and if smaller, TL and TH are made larger so that it can be easily determined as a still image.

The control threshold TV is set so as to substantially match the magnitude | V | of the motion vector V when the image is close to a still image.

The operation of the motion detector 21 having the above structure will be described. First, the operation of determining whether an image is a moving image or a still image will be described.

Video signal S1 input from the input terminal 11
And the video signal S11 output from the field memory 19
Are supplied to the interframe difference value calculation circuit 211. As a result, the inter-frame difference value FD is obtained for each pixel.

This interframe difference value FD is supplied to the image determination circuit 213. As a result, the absolute value | FD | of the inter-frame difference value FD and the two threshold values TH and TL output from the threshold value generating circuit 212 are compared in size, and based on the comparison result, the image is a moving image or a still image. Is determined.

This state is shown in FIG. As shown in the figure,
If the absolute value | FD | is smaller than the lower threshold value TL (| F
When D | ≦ TL), the input video is determined to be a still image. As a result, in this case, the inter-field / intra-field adaptive line interpolation coefficient γ is set to 1 (γ = 1). As a result, in the switching unit 17, the inter-field line interpolation unit 15
Only the video signal S7 output from is selected as the video signal S10.

If the absolute value | FD | is larger than the upper threshold value TH (| FD | ≧ TL), the image is determined to be a still image. As a result, in this case, the inter-field / intra-field adaptive line interpolation coefficient γ is set to 0 (γ
= 0). As a result, the switching unit 17 selects only the video signal S9 output from the intra-field line interpolating unit 16 as the video signal S10.

Further, the absolute value | FD | is equal to the thresholds TL and TH.
If (TL <| FD | <TH), it is determined that the image is neither a complete still image nor a complete moving image. As a result, in this case, the inter-field / intra-field adaptive line interpolation coefficient γ is set to a value according to the magnitude of | FD | (0 <γ <1). As a result, the switching unit 17 mixes the video signals S7 and S9 and outputs the mixed signal as the video signal S10.

The above is the operation of determining whether the image is a moving image or a still image. Next, the control operation of the motion detection thresholds TH and TL will be described.

The motion vector V detected by the motion vector detector 20 is supplied to the threshold control circuit 214. As a result, the magnitude | V | of the motion vector V is compared with a predetermined threshold value TV, and the threshold value T is calculated based on the comparison result.
H and TL are controlled.

This state is shown in FIG. As shown in the figure,
If | V | is larger than TV (| V | ≧ TV), the threshold T
L and TH are set to the smaller values TLL and THL, respectively. On the other hand, when | V | is smaller than TV (| V | <TV), the thresholds TL and TH are TL respectively.
It is set to a value TLH or THH larger than H and THH.

As a result, the judgment characteristic of the image judgment circuit 213 becomes C1 or C2 in FIG. here,
C1 is the characteristic when | V | is larger than TV, and C1
2 is a characteristic when | V | is smaller than TV.

As a result, when | V | is larger than TV,
In other words, when the motion vector V determines that the image is a moving image, the image is likely to be determined as a moving image. As a result, it is possible to reduce erroneous determination that a moving image is determined to be a still image, and thus it is possible to reduce the occurrence of comb distortion.

On the other hand, if | V | is smaller than TV, in other words, if the motion vector V determines that the image is a still image, the image is likely to be determined as a still image.

Here, | V | becomes smaller than TV in a horizontal edge portion having a small vertical correlation and in a region where the motion amount changes within one line or less. Therefore, in such a region, the image is likely to be determined as a still image, and the occurrence of the worm eating phenomenon is reduced.

Furthermore, the fact that | V | is smaller than TV means that the image is close to a still image. Therefore, even if the above-described control is performed, there is almost no malfunction that a moving image is determined to be a still image. As a result, even when the difference in brightness of the horizontal edge portion is small,
It is possible to reduce the occurrence of the worm-eaten phenomenon without inviting the occurrence of comb distortion.

According to this embodiment described in detail above, the following effects can be obtained.

(1) First, the magnitude of the motion vector V | V
When | becomes smaller than the predetermined threshold value TV, the motion detection threshold values TH and TL are increased to make it easier to determine that the image is a still image. be able to.

(2) Also, instead of the edge detection output,
Since the image determination operation is controlled based on the magnitude | V | of the motion vector V, the occurrence of the worm-eating phenomenon is reduced even at the horizontal edge portion where the luminance difference is small, without causing the comb distortion. can do.

(3) Further, the magnitude of the motion vector V |
When V | is larger than the predetermined threshold value TV, the motion detection threshold values TH and TL are set to the values THL and TLL that make it easier to determine an image as a moving image. Occurrence of can be reduced.

FIG. 8 is a block diagram showing the structure of the main part of the second embodiment of the present invention.

In the above embodiment, the case where the motion detection thresholds TH and TL are controlled based on the magnitude | V | of the motion vector V has been described. On the other hand, in this embodiment, the motion detection thresholds TH and TL are fixed and the interframe difference value FD is set.
Is controlled.

That is, in FIG. 8, reference numeral 215 is an interframe difference value calculation circuit for calculating the interframe difference value FD. The interframe difference value calculation circuit 215 is configured as a gain variable circuit. 216 is the motion vector V
The gain control circuit controls the gain G of the inter-frame difference value calculation circuit 215 based on the magnitude | V |

In such a configuration, the gain G of the inter-frame difference value calculation circuit 215 is a small value when the magnitude | V | of the motion vector V is larger than the control threshold TV, as shown in FIG. GL is set, and when it is small, it is set to a large value GH.

Accordingly, the magnitude of the motion vector V | V
When | is large, the image determination operation is controlled so that the image is easily determined to be a moving image, and when it is small, it is easily determined to be a still image.

Even in such a configuration, the image determination operation is controlled on the basis of the magnitude | V | of the motion vector V, so that the same effect as that of the previous embodiment can be obtained.

The two embodiments of the present invention have been described above in detail, but the present invention is not limited to the above-mentioned embodiments.

(1) For example, in the above-described first embodiment, when the motion detection thresholds TH and TL are controlled based on the magnitude | V | of the motion vector V, when these are controlled by the same amount. Explained. However, in the present invention, each control amount may be different.

(2) Further, in the above-mentioned first embodiment,
The case of controlling the threshold value for motion detection based on the magnitude | V | of the motion vector V and the case of controlling the two threshold values TH and TL have been described. However, the present invention may control only one of the thresholds.

(3) Further, in the above-mentioned first embodiment,
The case where the motion detection thresholds TH and TL are controlled stepwise has been described. However, in the present invention, these may be continuously controlled, or the staircase control and the continuous control may be appropriately combined. This also applies to the case where the gain G is controlled in the second embodiment.

(4) Further, in the above-described first and second embodiments, the case where the two thresholds TH and TL are used as the threshold for motion detection has been described. However, the present invention may use one threshold value or three or more threshold values.

(5) Further, in the above-mentioned first and second embodiments, the case where the image is judged based on the interframe difference value FD has been described. However, according to the present invention, if the image is judged based on the difference value of the video signals separated by N (N is an integer of 1 or more) fields, the interframe difference value F
The image may be determined based on a difference value other than D.

(6) In addition, in the above-described first and second embodiments, the present invention is applied to a motion detecting device for performing image determination for switching between interframe line interpolation and intraframe line interpolation. Explained. However, the present invention can also be applied to a motion detection device that performs image determination for other purposes.

(7) In addition to this, it is needless to say that the present invention can be variously modified without departing from the scope of the invention.

[0093]

As described above in detail, according to the present invention, even in the horizontal edge portion having a small brightness difference, it is possible to reduce the occurrence of the worm eating phenomenon without causing the comb distortion. A device can be provided.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a first exemplary embodiment of the present invention.

FIG. 2 is a diagram for explaining inter-field line interpolation and intra-field line interpolation.

FIG. 3 is a diagram showing comb distortion caused by inter-field line interpolation.

FIG. 4 is a diagram showing an insect-eating phenomenon that the present invention intends to reduce.

FIG. 5 is a characteristic diagram showing an image determination characteristic of the first embodiment.

FIG. 6 is a characteristic diagram showing a threshold control characteristic of the first embodiment.

FIG. 7 is a characteristic diagram showing changes in image determination characteristics in the first embodiment.

FIG. 8 is a block diagram showing a configuration of a main part of a second embodiment of the present invention.

FIG. 9 is a characteristic diagram showing a gain control characteristic of the second embodiment.

[Explanation of symbols]

11 ... Input terminal, 12 ... Motion compensation field interpolator, 1
3, 19, 22, 141, 142 ... Field memory,
14 ... Interframe interpolating unit, 15 ... Interfield line interpolating unit, 16 ... Field intraline interpolating unit, 17 ... Switching unit, 18 ... Output terminal, 20 ... Motion vector detecting unit, 21
... Motion detectors, 143, 144, 151, 152, 16
2, 163, 171, 172 ... Multiplier circuits, 145, 15
3, 164, 173 ... Addition circuit, 161, ... Line memory, 211, 215 ... Interframe difference value calculation circuit, 21
2 ... Threshold value generation circuit, 213 ... Image determination circuit, 214 ... Threshold value control circuit, 216 ... Gain control circuit.

Claims (1)

[Claims] 1. An image determination means for determining whether an image is a moving image or a still image by comparing an amplitude difference of video signals separated by N (N is an integer of 1 or more) fields with a predetermined threshold, and a moving image. If the magnitude of the motion vector indicating the magnitude and direction of the motion is small, it is easier to determine that the image is a still image. Based on the magnitude of this motion vector, a determination for controlling the image determination operation of the image determination means is performed. A motion detecting apparatus comprising: motion control means.