JPH0722383B2 - Motion vector detector - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a television broadcasting system, and more particularly to a method of correcting a moving part of an image such as system conversion or frequency band compression.

[Prior Art] When television broadcasting standard systems are different, in order to exchange broadcast programs with each other, conversion of the standard system is necessary, and a system conversion device for that purpose is used.

In the system conversion when the number of scanning lines or the number of fields is different in the standard system, the converted television image has a deterioration in resolution due to the conversion process or a moving image part has an unnatural and smooth motion. There is a problem that must be improved because there is no interference such as judder and image quality is greatly deteriorated.

Conventionally, in order to correct a moving image portion by this system conversion processing, a frame memory is used to obtain a motion vector from the correlation of the image signal level between the current frame and the immediately preceding frame, and The motion image was corrected.

Now, motion correction is performed using a motion vector, and a pattern matching method is used to detect this motion vector.

This is because, as shown in FIG. 4, when the object A moves at the speed of υ during one frame period, the motion vector υ coincides with the current frame image A ′ if the image A of the previous frame is shifted by υ. , Is detected by utilizing the principle that the difference in the luminance level of a pixel between frames is the smallest.

This is a determination method that uses the difference between pixel signal levels between frames to detect an optimum motion vector.

Therefore, with respect to the signal level of a pixel and each pixel signal level shifted by a predetermined vector type in order to detect in which direction the pixel has moved, the current frame and the previous frame, which are adjacent to each other, are detected. Calculate the difference between frames and use the sum of the absolute values to evaluate the correlation.

That is, since the vector having the smallest sum has the highest correlation, this is set as the correct motion vector for the image area.

However, when configuring the hardware using all the pixels of the current frame, the scale of the device becomes unnecessarily large, so in practice the pixels of the sub-sampled points called representative points are actually used. It is just arithmetic processing.

That is, as can be seen from FIG. 4, it is only the representative points that belong to the motion area that are effective in detecting the motion vector of the moving image A, and the other representative points are the still image areas. Therefore, if the representative point of the still image area is added to the sum calculation, the minimum value of the total sum of absolute values acts in the direction of the zero vector.

Therefore, it is unsuitable to add a representative point considered to be a still image area to the motion vector detection, and there is a disadvantage that it becomes useless.

Furthermore, the conventional method has a drawback that the accuracy of vector detection is deteriorated when a small-amplitude video signal has noise such as a fine image movement.

[Problems to be Solved by the Invention] Therefore, an object of the present invention is to add a frame difference signal level by using only a representative point of a moving image area in an image signal and the image signal to detect a motion vector. To improve the motion vector detection accuracy.

[Means for Solving the Problems] In order to achieve such an object, according to the present invention, a still image region and a moving image region are discriminated from a video input signal, and a representative frame belonging to a moving image region from a previous frame signal of the video signal. Only the points are sub-sampled, and each pixel signal level corresponding to this representative point is subtracted from the pixel signal level of a point belonging to the moving image area of the current frame, which is shifted from the representative point by a predetermined vector type. To do.

That is, the present invention includes a storage unit that stores one frame signal of a television image input signal, a moving image region detection unit that detects a moving image region of the television image input signal,
Representative point setting means for setting a representative point of a pixel belonging to the moving image area by sampling from the previous frame signal read from the storage means based on the detection signal of the moving image area detecting means, and a detection signal of the moving image area detecting means. On the basis of the above, a moving image extracting means for extracting only a signal belonging to a moving image area from the television image input signal of the current frame, and a pixel at a point shifted by a predetermined sample vector from each representative point set by the representative point setting means. And a subtraction means for performing a subtraction between the level of the video signal and the level of the video signal of the pixel at the point belonging to the moving image area of the current frame extracted from the moving image extracting means, and outputting the subtraction result as a frame difference. , The absolute value of the frame difference obtained from the subtracting means is accumulated by the number of representative points used for each sample vector. A frame difference addition means, characterized by comprising a minimum value detecting means for detecting a motion vector by detecting the minimum value of the cumulative value obtained from the frame difference addition means.

[Operation] According to the present invention, a sub-sampling pixel signal belonging to a moving image area of a television video signal can be used to determine a difference between signal levels between a previous frame and a current frame and a minimum addition value. It is possible to detect a motion vector with high accuracy.

Further, even for a moving image of a small image with a small amplitude, the correct motion vector can be detected without being caught by noise, so that a good image conversion without judder can be performed.

[Examples] Hereinafter, the present invention will be described in detail with reference to the drawings.

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

In the figure, 1 is a moving image region detection unit, 2 is a frame memory unit, 3 is a representative point setting unit, 4 is a switching circuit controlled by a moving image region detection signal, 5 is a subtractor, 6 is a frame difference addition unit, and 7 is It is a minimum value detection unit of the frame difference addition value.

Now, the video input signal is input from the input terminal T ₁ to the moving image area detection unit 1
Is input to the display area, and the still image area and the moving image area are discriminated here.

On the other hand, the previous frame signal of the video input signal output from the frame memory unit 2 is processed by the representative point setting unit 3 in the moving image area detecting unit 1.
The representative point, which is a pixel belonging to the sub-sampled moving image area, is set under the control of the detection signal from.

The difference between the signal level at a point shifted by a predetermined vector type (sample vector) from each of the representative points and the video signal level of the current frame output from the (1) side terminal of the switching circuit 4 is the subtraction circuit 5 But it can be stopped.

Therefore, both the representative point of the previous frame and the video signal of the current frame used here are limited to the moving image area by the control of the moving image area detecting unit 1.

For each sample vector, the frame difference addition unit 6 accumulates the absolute value of the frame difference by the number of representative points used.

Of the sample vectors, the one that is selected as the optimum motion vector based on the principle of the pattern matching method is the one that gives the smallest cumulative value.

This is detected by the minimum value detection unit 7, and the motion vector is output to the output terminal T ₂ .

When there is no detection signal from the moving image area detection unit 1, the switching circuit 4
Is connected to the terminal (2) side. Here, α is a constant value given to a pixel in the still image area. Generally, α = 0.

The portion surrounded by the dotted line in FIG. 1 indicates that the switching circuit 4 is controlled by the detection signal from the moving image area detecting unit 1, and only the signal belonging to the moving image area is extracted from the current frame signal.

FIGS. 2A to 2D are explanatory diagrams showing the operations of the moving image (A), the moving image area (B), the representative point (C), and the switching circuit output (D), respectively.

In FIG. 2 (A), the solid line represents the moving image of the motion vector υ, which is the current frame signal and the broken line is the previous frame signal.

The brightness level of the image is a ₀ and the amplitude is (a ₁ −a ₀ ).

FIG. 2 (B) shows the moving image area of FIG. 2 (A). The period from x _{2 to} x ₁ is a moving image area, which serves as a control signal for the representative point setting unit 3 and the switching circuit 4.

As shown in FIG. 2 (C), the representative points used for motion vector detection are only points d to g.

Also, controlled by the detection signal from the moving image area detection unit 1,
The video signal output from the switching circuit 4 has an amplitude value a ₁ as shown in FIG.

This is because the amplitude level a ₁ of the moving image signal is the constant reference amplitude value a ₀
It is shown that the amplitude of the motion image signal having the relationship of a ₁ > a ₀ becomes large because the _value is added to

Therefore, the frame difference value based on the sample vector other than the motion vector υ is increased.

FIGS. 3A and 3B show a comparison between the frame difference addition value according to the present invention and the conventional value, where FIG. 3A shows the conventional method and FIG. 3B shows the present invention.

As seen from this, the change due to each sample vector of the added value of the frame difference is steep in the dip state as shown in (B) of FIG. 3A, so that even in a noisy video signal, It can be seen that the accuracy of detecting the motion vector υ of a fine image is improved.

Note that, here, the control by the detection signal from the moving image area detecting unit 1 according to the present invention is limited to only the control of the switching circuit 4, and the control of the representative point setting unit 3 is not performed, and it is not limited to the moving image area. A large frame difference occurs in the motion vector υ at the points of pixels near the moving image area such as the representative points c and h in C),
This will reduce the accuracy of motion vector detection. Since the sub-sampling points are limited by newly adding the control of the present invention, there is an advantage that the scale of the apparatus can be reduced and the detection sensitivity can be further improved as compared with the conventional method. .

[Advantages of the Invention] In contrast to the conventional motion vector detection method, according to the present invention, by limiting the pixels to the moving image area and effectively using the brightness level, a moving image having a small amplitude and a small amplitude can be obtained. On the contrary, the accuracy of motion vector detection can be improved.

Furthermore, since it suffices to calculate the frame difference addition value only in the moving image area, there is an advantage that the size of the device can be reduced and the size can be reduced.

Although the present invention has been described above with reference to a standard system converter for television broadcasting, high-definition television (high-definition television system) which requires a wide frequency band is band-compressed by a channel having a limited narrow frequency band. Mu to transmit
By applying it to the se image transmission / reception system and the motion compensation type high efficiency band compression system, it is possible to reproduce a good image without judder.

[Brief description of drawings]

 FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is an explanatory diagram showing the operation of each part of the present invention, and FIG. 3 is a comparison diagram of the frame difference addition value between the present invention and the prior art, FIG. 4 is an explanatory diagram of the pattern matching method. 1 ... Moving image area detection unit, 2 ... Frame memory unit, 3 ... Representative point setting unit, 4 ... Switching circuit, 5 ... Subtractor, 6 ... Frame difference addition unit, 7 ... Frame difference addition value Minimum value detector.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Yuichi Ninomiya 1-10-11 Kinuta, Setagaya-ku, Tokyo Inside the broadcasting technology research institute of Japan Broadcasting Corporation (72) Taiji Nishizawa 1-10 Kinuta, Setagaya-ku, Tokyo No. 11 Inside the Japan Broadcasting Corporation Broadcasting Technology Laboratory

Claims (1)

[Claims] 1. A storage unit for storing one frame signal of a television image input signal, a moving image region detecting unit for detecting a moving image region of the television image input signal, and a detection signal of the moving image region detecting unit. A representative point setting means for setting a representative point of a pixel belonging to the moving image area from the previous frame signal read from the storage means by sampling; and a television image input of the current frame based on a detection signal of the moving image area detecting means. A moving picture extracting means for extracting only a signal belonging to a moving picture region from the signal; a video signal level of a pixel at a point shifted by a predetermined sample vector from each representative point set by the representative point setting means; and the moving picture extracting means The subtraction result is obtained by subtracting from the video signal level of the pixel at the point belonging to the video area of the current frame extracted from And a frame difference addition means for accumulating the absolute value of the frame difference obtained from the subtraction means by the number of representative points used for each sample vector, and the frame difference addition And a minimum value detecting means for detecting a motion vector by detecting a minimum cumulative value obtained from the means.