JPH025689A - Motion vector detector for picture - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application This invention relates to a motion vector detection device for television images.

2. Description of the Related Art A conventional image motion vector detection device is disclosed in, for example, Japanese Patent Laid-Open No. 61-201581. FIG. 8 shows a block diagram of this conventional image motion vector detection device.

In the figure, a representative point memory 2 stores the signals of pixels at the positions of some representative points among the image signals of one frame or one field input from the input terminal 1, and stores the signals of pixels at the positions of some representative points from the input terminal 1. When a frame or one field of image signals is input and supplied to the correlator 3, the signal of the stored representative point is supplied to the correlator 3. The correlator 3 calculates the absolute value 1ΔLl of the difference between the signal of the representative point and the signal of the pixel shifted (i, j) in the horizontal and vertical directions from the position of the representative point of the image signal.
(i.

j) and supplies it to the cumulative addition circuit 4. The cumulative addition circuit 4 calculates this 1ΔLl(i.

j) is cumulatively added to obtain a correlation value Σ1ΔLl (i, j) represented by the deviation ff1 (i, j), and this is supplied to the motion vector generation circuit 5. In the motion vector generation circuit 5,
Detect the deviation i (i, j) that minimizes this correlation value ΣIΔLl (i, j), generate a motion vector,
It is supplied to the output terminal 6.

However, with the above configuration, if the S/N ratio of the input image signal is poor, there is a high probability that an erroneous vector will be detected.

Furthermore, in order to reduce the probability of detecting the erroneous vector, the number of representative points within the detection area and the number of cumulative additions must be set to a certain value or more. Furthermore, even if the S/N ratio of the image signal is good, if the waveform of the image signal changes in a stepwise manner as shown in the screen (a) shown in FIG. 9 and the level wave (b) one line above it, As shown in the correlation value (C) in Figure 9, the correlation value ΣlΔL I U) takes a certain minimum value in a range 9 that is almost equal to the interval 8 between the representative points 7 and 7°, so the accuracy of vector detection deteriorates. , furthermore, motion vector detection becomes impossible. In order to prevent this, the interval between the representative points must be kept below a certain level, and therefore the number of representative points within the detection area and the number of cumulative additions must be above a certain level. Therefore, in the conventional method, the number of representative points in the detection area and the number of cumulative additions must be greater than or equal to the number of times of cumulative addition, so there is a certain limit to the simplification of hardware and the reduction of processing time. Furthermore, when the motion vector detection range is set within a certain value, and the pixel area within the range that deviates from each representative point within a certain value is set as the search area corresponding to that representative point, each Search area 10.1
0' do not overlap each other within the screen, and if there is a one-to-one correspondence between pixels and representative points in each search area, the apparatus shown in FIG. When one pixel signal of a frame or field that is manually entered later is input, the representative point memory 2 only has to output the signal of one representative point corresponding thereto, and the correlator 3 and cumulative adder 4 , only one operation is required for each, so the hardware can be greatly simplified. Also, the number of times each search area to, to' overlaps is large (as is true, the hardware becomes complicated.However, in order for each search area to not overlap, its size 11.11
' cannot be set to a distance of 8.8° or more between each representative point, so the range of movement f\tre that can be detected is limited thereby.

In the conventional method, as mentioned above, the interval between each representative point must be kept below a certain value, so if you try to simplify the hardware, the range of motion vectors that can be detected becomes narrow (or When trying to widen this range, there was a problem in that the hardware became complicated.

Third, due to flickering in the image signal or the activation of automatic exposure adjustment during shooting, the level of the signal waveform may change as shown in Figure 11(a).
When the signal changes as shown in (b), the absolute value 1ΔL l (i) of the difference between the signals from each representative point is 12.12' in (b) of the same figure.
The correlation value ΣlΔL l (i) obtained by cumulatively adding them becomes as shown in the same figure (C),
The minimum point 13 is a point different from the amount of movement of the image.

Therefore, an erroneous vector is detected.

Problems to be Solved by the Invention Therefore, in the conventional method, when the signal waveform level fluctuates due to flicker during shooting or the operation of an automatic exposure adjustment device, there is a high probability of erroneous vector detection. However, this method had the problem that motion vector detection was impossible.

However, regarding this issue, even in the conventional example, the above-mentioned 8th
In correlator 3 in the figure, instead of the absolute value of the difference between two signals,
The motion vector generation circuit 5 calculates the motion vector by multiplying the two signals, and changes the motion vector to the deviation amount (i, j) that gives the maximum value of the cumulative addition.
Solvable.

However, in this case, when configuring with hardware, the correlator 3
Therefore, a multiplication circuit is required. In addition, the number of bits of the digital signal of the correlator 3 is large, so the amount of calculations in the cumulative addition 4 and the motion vector generation circuit 5 after the correlator 3 is large, and the circuit scale thereof is also large. Therefore, there is a drawback that the circuit scale of the entire motion vector detection device is large (and complicated), and the processing time is also long.

In view of the above, the present invention aims, firstly, to provide a motion vector detection device that can accurately detect a motion vector even from an image signal with a poor S/N ratio; The motion vector cannot be detected accurately even from a bad image signal or an image signal whose signal waveform changes in a step-like manner, even if the intervals between representative points are wide, or the number of representative points and cumulative additions is small (but the motion vector cannot be detected accurately). Therefore, it is an object of the present invention to provide a motion vector detection device that can widen the range of motion vectors that can be detected without complicating the hardware, and further simplify the hardware and shorten the processing time. Thirdly, it is an object of the present invention to provide a motion vector detection device capable of accurately detecting a motion vector even when there is a fluctuation in the level of a signal waveform due to flicker or the like, without changing the circuit scale of the device.

Means for Solving the Problems The present invention provides a one-dimensional or two-dimensional low-pass filter that suppresses high-frequency components in one or both of vertical and horizontal directions of image information, and image information that has passed through the low-pass filter. An image motion vector detection device includes a circuit for determining a correlation value for a specific shift amount between two frames or two fields, and a circuit for detecting an image motion vector based on the correlation value. .

It also includes a high-pass filter that suppresses DC components and low-frequency components of image information, a circuit that calculates a correlation value for a specific shift amount of two frames or fields of image information that has passed through the high-pass filter, and the correlation value. This is an image motion vector detection device that includes a circuit that detects the motion vector of an image based on the .

By suppressing the high-frequency components of the working image information, the influence of noise can be suppressed, and motion vector detection becomes more accurate. Furthermore, by suppressing the high-frequency components of the image, even when a signal whose waveform changes stepwise as shown in FIG. 9(b) is input, the signal waveform that has passed the low-pass filter is As shown in figure (a), the signal becomes smooth, and the absolute value of the signal difference for each representative point is 1
The correlation value Σ1ΔL l (i) obtained by cumulatively adding up ΔL l (i) is shown in FIG. Here, the minimum point B of the correlation value is found at one point, and even if the intervals between the representative points are wide, it is possible to accurately detect the motion vector. Furthermore, by suppressing the DC component and low frequency component of the image, if there is a fluctuation in the signal waveform level due to flicker etc., for example as shown in Figure 10, if this is passed through a high pass filter, it will be possible to By taking the difference between separate signals, a waveform as shown in Fig. 5(a) is obtained, and by calculating the correlation value ΣjΔL l (i) from this, the correlation value ΣlΔL is obtained as shown in Fig. 5(C). (i) is obtained. When compared with FIG. 10C, it is clear that a more accurate motion vector can be detected than with the conventional method.

Embodiment FIG. 1 shows a block diagram of an image motion vector detection apparatus in a first embodiment of the present invention.

The operation will be explained below. First, image signals of at least two fields are input to the input terminal 1. The input image signal is supplied to a low-pass filter 14,
High frequency components in one or both of the horizontal and vertical directions are suppressed and supplied to the representative point memory 2 and the correlator 3. The representative point memory 2 stores, among these signals, signals of pixels at positions of some representative points. When the image signal of another field or frame is supplied to the correlator 3 through the low-pass filter 14, the signal of the representative point of the previous field or one frame is sent from the representative point memory 2 to the correlator 3. The correlator 3 takes the absolute value IΔL1 of the difference between the signal of the representative point and the signal of the pixel shifted (i, j) from the position of the representative point of another image signal, and cumulatively adds it. The rosewood addition circuit 4 cumulatively adds this absolute value 1ΔL1 for each deviation amount (i, j) for all representative points within the detection area, and calculates Σ1ΔLl (i, j
). This Σ1ΔL (i, j) is expressed as the deviation amount (i, j
) and is supplied to the motion vector generation circuit 5. The motion vector generation circuit 5 generates a correlation value Σ1ΔLl
Detect the deviation (i, j) that minimizes (i, j), and use this as the motion vector of the image in the detection area at the output terminal 6.
supply to.

This effect can suppress the influence of noise by suppressing high-frequency components of image information, making motion vector detection more accurate. By suppressing the high-frequency components of the image, even when a signal whose waveform changes in a stepwise manner as shown in FIG. As shown in a), the absolute value of the signal difference 1ΔL l (i) for each representative point is the same as that in (b), and the correlation value Σ1ΔL l (i) obtained by cumulatively adding them is the same. The result is as shown in Figure (C). Here, the minimum point 13 of the correlation value is found at one point, and even if the intervals between the representative points are wide, it is possible to accurately detect the motion vector.

FIG. 6 shows the first effect of the first embodiment of the present invention. The figure shows the presence and absence of a low-pass filter,
7 is a graph showing the relationship between the S/N ratio of an image and the average value of the error size of detected vectors. Here, as the above-mentioned low-pass filter, 11
An R type low pass filter was used. In the formula, Z and W represent a delay of one pixel in the horizontal direction and one line in the vertical direction, respectively.

; α=3/4. β-1/2: α=15/16. β-7/8 Further, the number of representative points within the detection area and the number of cumulative additions at this time are each 16. Conventionally, 20,010 or more are required, but in this embodiment, the number is significantly reduced. From Figure 6, compared to the conventional example that does not use a low-pass filter,
In the example of the present invention, the error in the detected hector is clearly small (and the motion vector can be detected more accurately, which has a large effect.Also, although a 11R type low-pass filter was used here, other low-pass filters may be used. .

FIG. 2 shows a block diagram of an image motion vector detection device according to a second embodiment of the present invention. Second
In the figure, 1 is an image signal input terminal, 15 is a high-pass filter, and the others are the same as in FIG.

The operation of the image motion vector detection device of this embodiment further configured as described above will be described below. First, an image signal input from an input terminal is supplied to a high-pass filter 15. Ahipass filter 15 suppresses DC components and low frequency components of the image signal, and supplies its output to representative point memory 2 and correlator 3. The following operations are the same as in the first embodiment. This effect will be discussed later.

FIG. 3 shows a block diagram of an image motion vector detection apparatus according to a third embodiment, which is a combination of the first embodiment and the second embodiment of the present invention. In FIG. 3, 16 is a band pass filter, which is the same as in FIG. Here, the filter 16 suppresses high-frequency components in one or both of the horizontal and vertical directions of the image, and at the same time suppresses DC components and low-frequency components.

The operation is as follows: 1. This is similar to the second embodiment.

This second. The effect of the third embodiment is that by suppressing the DC component and low-frequency component of the image, when there is a fluctuation in the signal waveform level due to flicker etc., for example, in the case shown in FIG. 4(a), this can be suppressed. For example, by taking the difference between signals separated by a certain number of pixels in the horizontal direction, the waveform shown in Figure 5 (a) is obtained, and from this, the waveform shown in Figure 5 (b) is obtained. Absolute value 1ΔL for each representative point
When l (i) is obtained, the cumulatively added correlation value Σ1ΔL l (i) is obtained as shown in FIG. When compared with FIG. 9(C), it is clear that a more accurate motion vector can be detected than in the conventional method.

This second. An example of the effect of the third embodiment will be explained using FIG. 7. FIG. 7 is a graph of the image signal flicker and the average value of vector error when a motion vector is detected from this signal.

Results are also shown for images with relatively strong contrast and images with weak contrast. HI in the figure) What is shown without F is shown in the above example, (α, β') = (15/16.771
8) is a motion vector detection device equipped with a low-pass filter. Also, those indicated as having HPF are equipped with a high-pass filter H'(z)-1-z 6 as shown in the following formula in addition to the above-mentioned low-pass filter, and therefore the band-pass filter is composed of a low-pass filter and a high-pass filter. This is the motion vector detection device of the third embodiment.

Comparing the two, a motion vector detection device without a high-pass filter is inaccurate or impossible to detect motion vectors even from an image signal with 1 to 4% flicker, whereas a motion vector detection device with a high-pass filter The motion vector detection device can accurately detect a motion vector even from an image signal with 16 to 50% flicker.
The effect is huge. Further, here, a high-pass filter that takes the difference between signals at 16 pixels apart in the horizontal direction as shown in the above equation is used, but other high-pass filters may be used.

Here, an example was shown in which one motion vector is detected from the correlation value between two frames or fields, but the screen is divided into several detection areas, the motion vector is determined for each detection area, and they are determined and averaged. The present invention can also be applied to a motion vector detection device that calculates a motion vector for the entire screen by doing this, or further reduces a vector detection error by temporally weighted averaging a plurality of motion vectors obtained for several fields. can be applied and its effects are large.

Furthermore, the two frames or fields referred to herein do not necessarily have to be temporally consecutive, and may be separated by one or several frames or one or several fields.

As described in detail, according to the present invention, firstly, motion vectors can be accurately detected even from image signals containing noise. Second, even from an image signal containing noise or an image signal whose waveform changes stepwise, motion vectors can be accurately detected even with a very small number of representative points and cumulative addition.

Therefore, if the present invention enables motion vector detection using 16 representative points, which conventionally required 256 representative points, the capacity of the representative point memory and the number of calculations can be reduced by 1/2.
8 is enough. In addition, since the interval between representative points can be made much wider than in conventional methods, it is possible to increase the size of the search area without overlapping each search area (or reducing the number of times they overlap each other). Therefore, it is possible to detect motion vectors in a wider range without complicating the hardware compared to conventional methods.Secondly, even when the level of the signal waveform fluctuates due to flicker etc. , it becomes possible to accurately detect motion vectors.

Therefore, the ability of the image motion vector detection device can be greatly improved and at the same time the cost can be extremely reduced, and the practical effects of the present invention are great.

[Brief explanation of the drawing]

1, 2, and 3 are relationship diagrams for comparing image motion vector detection devices according to different embodiments of the present invention, and FIG. 8 is a block diagram of a conventional image motion vector detection device. Figure 9 is a route map and operation waveform diagram for explaining the operation of the conventional example, Figure 10 is a diagram schematically showing the relationship between each search area and representative points, and Figure 11 is the operation of the conventional example. FIG. 2... Representative point memory, 3... Correlator,
4... Cumulative sum unit, 5... Motion vector generation circuit, 14... Low pass filter, 1
5...High-pass filter j6...Name of band-pass filter agent Patent attorney Toshio Nakao
Figure 12 and i) (C) (b) Figure 1? Fig. 10 IQ' 7 Sotsuo] ノ/ Fig.

Claims (6)

[Claims] (1) A one-dimensional or two-dimensional low-pass filter that suppresses high-frequency components in one or both of the vertical and horizontal directions of image information, and two frames or two frames of image information that have passed through the low-pass filter. An image motion vector detection device comprising: a circuit for determining a correlation value for a specific amount of deviation of a field; and a circuit for detecting an image motion vector based on the correlation value. (2) a high-pass filter that suppresses DC components and low-frequency components of image information; and a circuit that calculates a correlation value for a specific shift amount of two frames or fields of image information that have passed through the high-pass filter; An image motion vector detection device comprising a circuit for detecting an image motion vector based on the correlation value. (3) A band-pass filter that suppresses high-frequency components, DC components, and low-frequency components in one or both of the vertical and horizontal directions of image information, and two frames of image information that have passed through the band-pass filter. Alternatively, an image motion vector detection device comprising: a circuit for determining a correlation value for a specific shift amount of a field; and a circuit for detecting an image motion vector based on the correlation value. (4) A circuit that calculates a correlation value for a specific shift amount of two frames or two fields of image information defines one or more detection areas from the screen, and further performs multiple search operations within the detection area. A region is determined, and a signal of one pixel at the position of a representative point of the search region of one frame or field;
Find the absolute value of the difference between the signal of the pixel at a specific deviation from the position of the representative point of the corresponding search area in the other frame or field, and accumulate it for some or all search areas in the detection area. 4. The motion vector detection device for an image according to claim 1, further comprising performing addition and using the cumulatively added value as a correlation value for a specific shift amount of the detection area. (5) The cutoff frequency f_c of the high frequency range in the vertical or horizontal direction is f_cf_s/2 when the sipling frequency f_s of the pixel in the vertical or horizontal direction is set. motion vector detection device for images. (6) Image motion vector detection according to claim 2 or 3, wherein the filter that suppresses the DC component and low-frequency component of the image information takes the difference between signals of adjacent pixels or pixels separated by a certain amount. Device.