JPH08251474A - Motion vector detector, motion vector detection method, image shake correction device, image tracking device and image pickup device - Google Patents

Abstract

PURPOSE: To conduct image shake correction and image tracking processing efficiently by allowing the device detecting the motion vector of an image to detect the motion vector of an entire image with high accuracy even when plural moving objects are in existence in the image. CONSTITUTION: A motion vector detection circuit 13 detects the motion vector of an image at plural block positions on a screen pattern and a reliability discrimination means 14 discriminates whether or not a detected motion vector is valid. Furthermore, the device is provided with a pattern division means 4 dividing the screen pattern into plural areas each having a different motion vector and with an area selection means 5 selecting any of the area divisions according to a prescribed condition, and a motion vector discriminated to be valid by the discrimination means 14 is fed to the means 4, 5. Then a motion vector decision means 15 decides the motion vector of the entire image from the motion vector in the area selected by the area selection means 5 and the image shake correction and image tracking processing is conducted by using the motion vector.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a motion vector detecting device for detecting a motion vector of an image, a motion vector detecting method, an image blur correcting device using the motion vector detecting device,
The present invention relates to an image tracking device and an imaging device.

[0002]

2. Description of the Related Art Conventionally, as a means for detecting a motion vector of an image, a correlation method and a block matching method based on a correlation calculation are known.

In the block matching method, an input image signal is divided into a plurality of blocks of appropriate size (for example, 8 pixels ×).
8 lines), calculate the difference between the previous field (or frame) and a certain range of pixels in block units, and calculate the block of the previous field (or frame) where the sum of the absolute values of this difference is the smallest. Is a way to explore.
The relative shift between the screens indicates the motion vector of that block.

Regarding the matching calculation, Morio Onoue et al., Information Processing Vol.17, No.7, p.634-640 July 1
It is discussed in detail in 976.

FIG. 13 is a block diagram showing the structure of a conventional motion vector detecting apparatus using the block matching method. The structure and operation will be briefly described below with reference to this figure.

When an image signal for which a motion vector is to be detected is applied to the spatial frequency filter 10, the filter 10
Extracts a spatial frequency component useful for motion vector detection from the image signal. That is, the low spatial frequency component and the high spatial frequency component of the image signal are removed.

The image signal passed through the filter 10 is
It is applied to the correlation calculation circuit 11 and the memory 12 provided as a 1-field period delay means. Correlation calculation circuit 11
Further, the image signal of the previous field is applied to the memory 12. Then, the correlation calculation circuit 11 performs the correlation calculation of the current field and the previous field in block units according to the block matching method, and outputs the result to the motion vector detection circuit 13.

The motion vector detection circuit 13 uses the correlation value from the correlation calculation circuit 11 to detect a motion vector in block units. Specifically, the block in the previous field having the smallest correlation value is searched for, and the relative shift is used as the motion vector.

The motion vector detection circuit 13 sends the detected motion vector in block units to the reliability judgment means 14. The reliability determination means 14 evaluates the reliability of the motion vector of each block and outputs a highly reliable motion vector to the motion vector determination means 15.

The motion vector decision means 15 decides the entire motion vector from the motion vectors sent from the reliability decision means 14. This entire motion vector is obtained as the average value or median value of highly reliable motion vectors. Then, the motion vector of this area is output as the motion vector of the entire image.

[0011]

However, in the above-mentioned conventional example, the processing when there are a plurality of moving objects in the image is insufficient, and there are the following problems.

That is, if the average value of the detected motion vectors is used as the entire motion vector, the motions of a plurality of moving objects in the image are mixed. For example, when the image blur correction process is performed using the average value, the screen moves like being pulled by a moving object.

If the median value of the detected motion vectors is used as the whole motion vector, the reliability of the motion vector will differ depending on the size and number of moving objects in the image. For example, when image blur correction is performed using the median value, if there are many moving objects, the image stabilization target changes at any time, and appropriate image blur correction cannot be performed.

The present invention has been made by paying attention to the above problems, and it is possible to accurately detect the motion vector of the entire image even when there are a plurality of moving objects in the image. An object of the present invention is to provide a detection device and a detection method thereof, an image blur correction device using the same, an image tracking device, and an imaging device equipped with these devices.

[0015]

A motion vector detecting device according to the present invention is configured as follows.

(1) A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions on a screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and An area selection means for selecting one according to a predetermined condition and a motion vector determination means for determining a motion vector representing the motion of the entire image from the motion vector of the selected area are provided.

(2) In the device of (1) above, the screen division means changes the division accuracy according to the magnitude of the motion vector at each position on the screen.

(3) In the device of (1) or (2), the area selecting means selects the area having the largest range from the plurality of areas.

(4) In the above device (1) or (2), the area selecting means selects the area having the smallest motion vector among the plurality of areas.

(5) In the apparatus of (1) or (2) above, the area selecting means selects the area having the largest range overlapping the previously selected area among the plurality of areas.

(6) In the apparatus of (1) or (2) above, the area selecting means overlaps the area having the largest range, the area having the smallest motion vector, and the area selected the previous time among the plurality of areas. Any one of the areas having the largest range is selected.

The image blur correction device according to the present invention is configured as follows.

(7) A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions on the screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and The motion vector detecting device comprises a region selecting means for selecting one and a motion vector determining means for determining a motion vector representative of the motion of the entire image from the motion vector of the selected region.

(8) In the device of (7) above, the screen division means of the motion vector detection device changes the division accuracy according to the magnitude of the motion vector at each position on the screen.

(9) In the above device (7) or (8), the area selecting means of the motion vector detecting device selects the area having the largest range, the area having the smallest motion vector, and the previous area among the plurality of areas. Any one of the areas having the largest overlapping area with the selected area is selected.

The image tracking device according to the present invention is constructed as follows.

(10) A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions on the screen, a screen dividing means for dividing the screen into a plurality of regions each having a different motion vector, and the divided regions. The motion vector detecting device comprises a region selecting means for selecting one and a motion vector determining means for determining a motion vector representative of the motion of the entire image from the motion vector of the selected region.

(11) In the device of (10) above, the screen dividing means of the motion vector detecting device changes the division accuracy according to the magnitude of the motion vector at each position on the screen.

(12) In the device of (10) or (11) above, the area selecting means of the motion vector detecting device is
Among the plurality of areas, one of the area having the largest range, the area having the smallest motion vector, and the area having the largest range overlapping with the previously selected area is selected.

The image pickup device according to the present invention is configured as follows.

(13) A motion vector detection circuit for detecting motion vectors of an image at a plurality of positions on the screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and An image blur correction device having a motion vector detection device including a region selection unit that selects one and a motion vector determination unit that determines a motion vector representing the motion of the entire image from the motion vector of the selected region. I did it.

(14) In the apparatus of (13) above, the screen division means of the image blur correction device changes the division accuracy according to the magnitude of the motion vector at each position on the screen.

(15) In the apparatus of (13) or (14), the area selection means of the image blur correction apparatus selects the area having the largest range, the area having the smallest motion vector, and the previous area among the plurality of areas. Any one of the areas having the largest overlapping area with the selected area is selected.

The image pickup device according to the present invention is constructed as follows.

(16) A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions on the screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and An image tracking device having a motion vector detection device including a region selection means for selecting one and a motion vector determination means for determining a motion vector representative of the motion of the entire image from the motion vector of the selected region. I chose

(17) In the above device (16), the screen division means of the image tracking device changes the division accuracy according to the magnitude of the motion vector at each position on the screen.

(18) In the apparatus of (16) or (17), the area selecting means of the image tracking apparatus selects the area having the largest range, the area having the smallest motion vector, and the previous selection from the plurality of areas. One of the regions having the largest range overlapping the selected region is selected.

(19) The motion vector detecting method according to the present invention detects the motion vector of the image at a plurality of positions on the screen, divides the screen into a plurality of regions each having a different motion vector, and divides the divided region. One area is selected from a plurality of areas according to a predetermined condition, and a motion vector representing the motion of the entire image is detected from the motion vector of the selected area.

(20) In the detection method of (19), the area having the largest range, the area having the smallest motion vector, and the area overlapping the previously selected area are selected from the plurality of areas of the divided screen. Is to select one of the largest areas.

[0040]

In the motion vector detecting apparatus and the detecting method thereof according to the present invention, the screen is divided into a plurality of areas each having a different motion vector, and one area is selected from the areas according to a predetermined condition. Since the motion vector of the entire image is determined from, the motion vector of the entire image can be accurately detected.

Since the image blur correction device of the present invention is equipped with the motion detection device having the above-described configuration, it is possible to perform accurate image blur correction.

Since the image tracking device of the present invention is equipped with the motion detecting device having the above-mentioned configuration, it is possible to perform the image tracking process with high accuracy.

Further, since the image pickup apparatus of the present invention has the image blur correction apparatus or the image tracking apparatus provided with the above motion vector detection apparatus, high-performance image pickup can be performed.

[0044]

Embodiments of the present invention will be described below with reference to the drawings.

(Embodiment of Motion Vector Detecting Device) FIG. 1 is a block diagram showing a schematic structure of a motion vector detecting device according to the present invention, and the same reference numerals as those in FIG. 13 denote the same constituent elements.

The image signal which is the object of motion vector detection is first applied to the spatial frequency filter 10. The filter 10 extracts a spatial frequency component useful for motion vector detection from the image signal. That is, the low spatial frequency component and the high spatial frequency component of the image signal are removed.

The image signal passed through the filter 10 is
It is applied to the correlation calculation circuit 11 and the memory 12 provided as a 1-field period delay means. Correlation calculation circuit 11
Further, the image signal of the previous field is applied from the memory 12. Then, the correlation calculation circuit 11 performs the correlation calculation of the current field and the previous field in block units according to the block matching method described above, and outputs the result to the motion vector detection circuit 13.

The motion vector detection circuit 13 detects the motion vector of the image at a plurality of block positions on the screen, and uses the correlation value from the correlation calculation circuit 11 to detect the motion vector in block units. . Specifically, the block in the previous field having the smallest correlation value is searched for, and the relative shift is used as the motion vector.

The motion vector detection circuit 13 sends the detected motion vector in block units to the reliability judgment means 14. The reliability determination means 14 evaluates the reliability of the motion vector of each block and determines whether it is valid or invalid. Then, the highly reliable motion vector is output to the screen dividing unit 4.

The screen division means 4 divides the screen into a plurality of areas having different motion vectors based on the motion vector sent from the reliability determination means 14, and the division result is sent to the area selection means 5. Output.

The area selecting means 5 selects one area from the plurality of areas divided by the screen dividing means 4 according to a predetermined condition, and outputs the selection result to the motion vector determining means 15. .

Then, the motion vector determining means 15 determines the motion vector of the entire area from the motion vector of the area selected by the area selecting means 5. This whole motion vector is obtained as the average value or median value of highly reliable motion vectors, and the motion vector of this region is output as the motion vector of the entire image.

In this way, the screen is divided into a plurality of areas each having a different motion vector, one area is selected from the areas according to a predetermined condition, and a motion vector representing the entire image is determined from the motion vectors of the area. Therefore, even when there are a plurality of moving objects in the image, the motion vector of the entire image can be accurately detected.

(Embodiment of Image Blur Correction Device and Image Tracking Device) FIG. 2 is a block diagram showing a schematic configuration of an image blur correction device and an image tracking device equipped with the above motion vector detection device. The device and the image tracking device are represented by the same configuration block.

In FIG. 2, the input chroma signal is stored in the memory 2
0, and the input luminance signal is added to the memory 21 and the motion vector detecting device 1 having the configuration of FIG. Then, the memory 20 temporarily stores the input chroma signal, and the memory 21 temporarily stores the input luminance signal.

The motion vector detecting device 1 detects the motion vector of the entire image from the input luminance signal as described above, and sends the detected motion vector to the integrating means 22.

The integrating means integrates the detected motion vector, and the memory reading means 23 controls the read parts of the memory 20 and the memory 21 according to the integrated motion vector. This allows the chroma signal from the memory 20,
Luminance signals are transmitted from the memory 21.

The image blur correction device and the image tracking device configured as described above can perform image blur correction and image tracking processing with high accuracy, respectively, and can realize a high-performance device.

(Embodiment of Image Pickup Apparatus) FIG. 3 is a block diagram showing the schematic arrangement of an image pickup apparatus having the above-described image blur correction apparatus or image tracking apparatus. The configuration and operation will be described below.

In FIG. 3, the subject image is the image pickup lens 3
After passing through 0, it is converted into an electric signal by the CCD 31 which is an image pickup device. The output of the CCD 31 passes through a sample hold (S / H) circuit 32 and automatic gain control (AGC).
It is sent to the circuit 33.

The analog output of the AGC circuit 33 is the A / D
It is converted into a digital signal by the converter 34 and input to the Y / C separation circuit 35 for the luminance signal (Y) and the color signal (C). The Y / C separation circuit 35 includes two 1H delay lines and an adder that adds an input signal and a 2H delay signal, and the output of the adder is a C process circuit 36 that is a color signal processing circuit.
In addition, the 1H delay signal is applied to the Y process circuit 37 which is a luminance signal processing circuit.

The C process circuit 36 is the Y / C separation circuit 3
A chroma signal is generated from the color signal output from the Y / C circuit 5, and the Y process circuit 37 performs edge enhancement, gamma correction, and the like on the luminance signal output from the Y / C separation circuit 35.

The chroma signal output from the C process circuit 36 and the luminance signal output from the Y process circuit 37 are input to the image blur correction device or the image tracking device 2 having the configuration of FIG. In the case of the image blur correction device, the blur-corrected chroma signal and luminance signal are output from here. In the case of an image tracking device, the chroma signal and the luminance signal subjected to the tracking processing are output from here.

The chroma signal output from the image blur correction device or the image tracking device 2 is converted into an analog signal by the D / A converter 38 and output. The brightness signal output from the image blur correction device or the image tracking device 2 is converted into an analog signal by the D / A converter 39 and output.

In this way, a video signal whose image blur has been corrected or a video signal which has undergone image tracking processing can be obtained. By providing this image blur correction device or image tracking device 2, a high-performance image pickup device can be obtained. it can.

Next, details of the screen dividing means 4 of FIG. 1 will be described with reference to the flow charts of FIGS.

(Embodiment 1 of the screen dividing means) FIG. 4 is a flow chart showing an example of the operation of the screen dividing means 4 in FIG.

First, the screen division means 4 receives and receives the motion vector judged to be valid by the reliability judgment means 14 (step 401). Next, the magnitude of the motion vector is compared with a plurality of thresholds. here,
The X component of the motion vector is compared with a threshold value A and a threshold value -A (A is a constant) (steps 402 and 403) to set the motion vector to a set larger than A (A), a set not smaller than -A and not larger than A (0), It is classified into three sets, which are smaller than -A (-A) (steps 404, 405, and 406).

Similarly, if the Y component is also classified into three sets, nine sets (sets (-A, A) and sets (0) represented two-dimensionally by the combination of X and Y as shown in FIG. , A), set (A, A), set (-A,
0), set (0, 0), set (A, 0), set (-
A, −A), set (0, −A), set (A, −A)) can be divided into motion vectors. For example, the motion vector indicated by the arrow in FIG. 10 is classified into the set (A, A).

At this time, since the detection block of each motion vector is known, if the detection block corresponding to the set of motion vectors is classified, the image can be divided into nine regions having different motions. After the division, the area information of each detection block is stored in the memory (step 407).

The above operation will be specifically described with reference to FIG. FIG. 11A schematically shows the input image, and there is one vehicle moving to the left and right of the screen. Further, FIG. 11B shows the motion vector detected by each detection block in the input image.

In the detection block where a car exists, a large motion vector is detected in the X direction (horizontal direction of the screen),
In the background detection block, a small motion vector is detected in the Y direction (vertical direction of the screen). Here, for the sake of simplicity, if all the vectors have high reliability and the reliability determination means 14 determines that they are valid, if the threshold value A is set to be smaller than the moving speed of the vehicle, the operation of the above-mentioned flowchart is performed. Accordingly, the motion vector of the vehicle moving to the left is classified into a set (-A), the motion vector of the background is classified into a set (0), and the motion vector of the vehicle moving to the right is classified into a set (A).

Therefore, the screen is divided into three areas (-A), area (0), and area (A) as shown in FIG. 11 (c).

Second Embodiment of Screen Dividing Means Although the threshold value is fixed in the above example, if the number of threshold values or the value is changed,
You can control the accuracy of screen division. Figure 5
Is a flow chart showing the operation of the screen division means 4 in which the division accuracy is changed according to the magnitude of the motion vector at each block position on the screen, and the steps with the same reference numerals as in FIG. 4 show the same processing contents. .

The difference from the operation of the first embodiment of the screen dividing means 4 in FIG. 4 is that a plurality of threshold value comparison routines are prepared so that they can be adaptively selected. In a plurality of comparison routines, different numbers of threshold values and different threshold values are prepared in advance, and the screens can be divided with different accuracy by passing through the respective routines.

For example, the number of regions divided in the previous field is discriminated (step 408), and if the number is smaller, the comparison routine is selected so as to divide the screen with a finer precision (step 411), and the division is performed in reverse. If the number of regions is large, the comparison routine is selected so that the screen division is performed with coarser accuracy (step 409), and if it is intermediate, the intermediate comparison routine is selected (step 410). Thereby, the motion vector of a certain area can be detected with higher accuracy.

Next, the details of the area selecting means 5 of FIG. 1 will be described with reference to the flow charts of FIGS.

(Embodiment 1 of Area Selecting Means) FIG. 6 is a flow chart showing an example of the operation of the area selecting means 5 in FIG.

The area information stored by the screen dividing means 4 is read (step 501), and the number of detection blocks constituting each area after the screen division is measured (step 502). Next, the measured numbers are compared (step 503) and the region with the largest number is selected (step 504). Finally, the motion vector of the selected area is transmitted (step 505).

The above operation will be specifically described with reference to FIG. FIG. 11D shows the area information after the input image of FIG. 11A is divided according to the operation of the first embodiment of the screen dividing unit 4 described above. This area information is information that distinguishes areas for each detection block. In this example,
The screen is divided into three areas. The number of each area is 13 in the area (-A), 49 in the area (0), and 8 in the area (A). (0) is selected.

As described above, by detecting the motion vector of the largest object in the image, when the camera is fixed in the TV conference system or the like and the image is occupied by almost one person, Therefore, effective image tracking processing can be performed.

(Second Embodiment of Area Selecting Means) FIG. 7 is a flow chart showing another example of the operation of the area selecting means 5.

The area information stored by the screen dividing means 4 is read (step 501), the magnitude of the movement in each area is compared, and the step (506) is performed to find the area with the smallest movement. Here, since the screen is divided by the size of the motion vector by the screen dividing means 4, the area having the smallest absolute value in the area information is selected (step 507). Then, the motion vector of the finally selected area is transmitted (step 505).

The above operation will be specifically described with reference to FIG. FIG. 11D shows the area information after the input image of FIG. 11A is divided according to the operation of the first embodiment of the screen dividing unit 4 described above. This area information is information that distinguishes areas for each detection block. In this example,
The screen is divided into three areas, with 0 in each area.
Has the smallest absolute value, the region (0) representing the background is selected.

As described above, by detecting the motion vector of the smallest object in the image, the image captured by the video camera, for example, is captured while keeping the subject stationary so that the image blur is very effective. Corrections can be made.

(Third Embodiment of Area Selecting Means) FIG. 8 is a flow chart showing another example of the operation of the area selecting means 5.

The area information stored by the screen dividing means 4 in the current field is read out (step 50).
1) Read the previously selected area information stored by the area selecting means 5 at the previous field processing (step 50).
8). Then, with respect to the region selected in the previous field, the number of regions divided in the current field, that is, the number of detected blocks in each region in the previous selected region is measured (step 509).

Next, the measured numbers are compared (step 503) and the region with the largest number is selected (step 5).
04). The information on the position data of the selected area is stored in the memory for comparison at the next area selection (step 510). Finally, the motion vector of the selected area is transmitted (step 505).

The above operation will be specifically described with reference to FIG. 12A and 12B show an input image of two consecutive fields. In addition, FIG.
The selected area information stored by the area selection unit 5 during the image processing of 2 (a) is shown, and the area shaded in the drawing is the selected area.

FIG. 12D shows the area information after the input image of FIG. 12B is divided according to the operation of the first embodiment of the screen dividing means 4 described above. Then, in the area information of (d) of FIG. 12 in the selected area of (c) of FIG. 12, the area (-A) that occupies the maximum number is selected as shown in (e) of FIG. Therefore, the selected area information of the image of FIG. 12B becomes as shown in FIG. 12F, and this information is saved in preparation for the next field processing.

As described above, by continuously detecting the motion vector of a certain object in an image, it is possible to detect the motion vector of an object regardless of the change in the size or speed of the object. Therefore, it is possible to perform very effective image blur correction and image tracking processing.

Fourth Embodiment of Area Selecting Means FIG. 9 is a flow chart showing another example of the operation of the area selecting means 5.

The area information saved by the screen dividing means 4 in the current field and the selected area information saved by the area selecting means 5 during the previous field processing are read (steps 501 and 508). If there is a region selected in the previous field (step 511), the number of regions divided by the current field in that region is measured (step 509). Next, the measured numbers are compared (step 503) and the region with the largest number is selected (step 504).

If there is no area selected in the previous field (step 511), the magnitude of the movement within each area is compared according to the area information stored by the screen dividing means 4 (step 506), and the largest movement is detected. A small area is searched for and selected (step 507). At this time, since the screen is divided by the size of the motion vector by the screen dividing means 4, the area having the smallest absolute value in the area information is selected. Then, the position data of the selected area is saved for comparison at the next area selection (step 5).
10). Finally, the motion vector of the selected area is transmitted (step 505).

[0095]

As described above, according to the present invention,
Since the screen is divided into multiple areas with different motion vectors, one area is selected according to the specified conditions, and the motion vector of the entire image is determined from the motion vector of that area. Even when the moving object exists, the image can be accurately obtained by dividing the image into a moving area where each moving object exists and a stationary area where there is no moving object.

By adaptively changing the division accuracy of the screen division means, the motion vector of a certain area can be obtained with higher accuracy.

By selecting the area of the motion vector having the largest range by the area selecting means, the camera used in, for example, the TV conference system is fixed, but the image is occupied by almost one person. In such cases,
Image tracking processing can be performed very effectively.

Further, by selecting the area of the motion vector having the smallest movement by the area selecting means, the photographer is aware that the object to be photographed does not move in the image photographed by the video camera. Can be done very effectively.

Further, by continuously detecting the motion vector of one object in the image in time, the motion vector of the object can be detected regardless of the change in the size or speed of the object, and the image blur correction can be performed. Processing and image tracking processing can be performed very effectively.

Further, it is possible to realize a high-performance image pickup apparatus capable of performing the above-described image blur correction processing or image tracking processing.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a motion vector detection device according to the present invention.

FIG. 2 is a block diagram showing the configurations of an image blur correction device and an image tracking device provided with a motion vector detection device.

FIG. 3 is a block diagram showing a configuration of an image pickup apparatus having an image blur correction device or an image tracking device.

FIG. 4 is a flowchart showing an example of the operation of the screen dividing means.

FIG. 5 is a flowchart showing another example of the operation of the screen dividing means.

FIG. 6 is a flowchart showing an example of the operation of the area selection means.

FIG. 7 is a flowchart showing another example of the operation of the area selection means.

FIG. 8 is a flowchart showing another example of the operation of the area selection means.

FIG. 9 is a flowchart showing another example of the operation of the area selection means.

FIG. 10 is an explanatory diagram showing classification of motion vectors.

FIG. 11 is an explanatory diagram showing specific operations of the screen dividing unit and the region selecting unit.

FIG. 12 is an explanatory diagram showing a specific operation of the area selection means.

FIG. 13 is a block diagram showing the configuration of a conventional motion vector detection device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 motion vector detection device 2 image blur correction device or image tracking device 4 screen division means 5 area selection means 11 correlation calculation circuit 13 motion vector detection circuit 14 reliability determination means 15 motion vector determination means 31 CCD

Claims (20)

[Claims] 1. A motion vector detection circuit for detecting motion vectors of an image at a plurality of positions in a screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and one of the divided regions. Motion vector detecting apparatus, which comprises: a region selecting unit that selects one of the two in accordance with a predetermined condition; and a motion vector determining unit that determines a motion vector representing the motion of the entire image from the motion vector of the selected region . 2. The motion vector detection device according to claim 1, wherein the screen division means changes the division accuracy according to the magnitude of the motion vector at each position on the screen. 3. The area selecting means selects the area having the largest range among the plurality of areas.
Alternatively, the motion vector detecting device described in 2. 4. The motion vector detecting apparatus according to claim 1, wherein the area selecting means selects an area having the smallest motion vector among the plurality of areas. 5. The motion vector detecting apparatus according to claim 1, wherein the area selecting means selects an area having the largest range of overlapping with the previously selected area among the plurality of areas. 6. The area selecting means selects one of a plurality of areas, that is, the area having the largest range, the area having the smallest motion vector, and the area having the largest range overlapping the previously selected area. The motion vector detecting device according to claim 1 or 2, characterized in that 7. A motion vector detection circuit for detecting motion vectors of an image at a plurality of positions in a screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and one of the divided regions. An image blur, which is provided with a motion vector detection device including a region selection unit that selects one of the two and a motion vector determination unit that determines a motion vector representative of the motion of the entire image from the motion vector of the selected region. Correction device. 8. The image blur correction device according to claim 7, wherein the screen division means of the motion vector detection device changes the division accuracy according to the magnitude of the motion vector at each position on the screen. 9. The area selecting means of the motion vector detecting device is any one of a plurality of areas having a largest range, an area having a smallest motion vector, and an area having a largest range overlapping the previously selected area. 9. The image blur correction device according to claim 7, wherein the image blur correction device is selected. 10. A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions in a screen, a screen dividing means for dividing the screen into a plurality of regions each having a different motion vector, and one of the divided regions. Image tracking, which is provided with a motion vector detection device including a region selection unit that selects one of the two and a motion vector determination unit that determines a motion vector representative of the motion of the entire image from the motion vector of the selected region. apparatus. 11. The image tracking device according to claim 10, wherein the screen division means of the motion vector detection device changes the division accuracy according to the magnitude of the motion vector at each position on the screen. 12. The area selecting means of the motion vector detecting device is any one of a plurality of areas having a largest range, an area having a smallest motion vector, and an area having a largest range overlapping with a previously selected area. The image tracking device according to claim 10 or 11, characterized in that 13. A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions on a screen, a screen dividing means for dividing the screen into a plurality of regions having different motion vectors, and one of the divided regions. An image blur correction device including a motion vector detection device including a region selection unit that selects one of the two and a motion vector determination unit that determines a motion vector representative of the motion of the entire image from the motion vector of the selected region. An imaging device characterized by. 14. The screen dividing means of the image blur correction device,
14. The image pickup apparatus according to claim 13, wherein the division accuracy is changed according to the magnitude of the motion vector at each position on the screen. 15. The area selection means of the image blur correction device comprises:
15. The one of a plurality of regions, which has a largest range, a region having a smallest motion vector, and a region which has a largest range overlapping with a previously selected region, are selected. Imaging device. 16. A motion vector detection circuit for detecting a motion vector of an image at a plurality of positions in a screen, a screen dividing means for dividing the screen into a plurality of regions each having a different motion vector, and one of the divided regions. An image tracking device having a motion vector detection device including a region selection unit that selects one of the two and a motion vector determination unit that determines a motion vector representing the motion of the entire image from the motion vector of the selected region. A characteristic imaging device. 17. The image pickup apparatus according to claim 16, wherein the screen division means of the image tracking device changes the division accuracy according to the magnitude of the motion vector at each position on the screen. 18. The area selecting means of the image tracking device selects one of a plurality of areas, the area having the largest range, the area having the smallest motion vector, and the area having the largest range overlapping the previously selected area. 18. The image pickup apparatus according to claim 16, wherein the image pickup apparatus is selected. 19. A motion vector of an image at a plurality of positions on a screen is detected, the screen is divided into a plurality of regions having different motion vectors, and a predetermined condition is selected from among the plurality of divided regions. A motion vector detection method characterized by selecting one area and detecting a motion vector representing the motion of the entire image from the motion vector of the selected area. 20. Selecting, from a plurality of divided screen areas, an area having a largest range, an area having a smallest motion vector, or an area having a largest range overlapping with a previously selected area. 20. The motion vector detecting method according to claim 19, further comprising: