JPH07322126A - Motion vector detecting circuit and subject tracking camera device using the detecting circuit - Google Patents

Abstract

PURPOSE:To track a subject with high accuracy even when the subject has a small size. CONSTITUTION:The image data on a subject are given to an image processing circuit from an image pickup element and also given to a motion vector detecting circuit as video signals. The motion vector detecting circuit stores the data on the key point of a precedent field in a key point memory 90 via a latch circuit 56. The data on the key point and the data on the present field of each detected block (d) received from the latch circuits 76a-76d are calculated by the correlators 102a-102d. Thus the differential absolute value is obtained. It is noticed here that the detected blocks are overlapping with each other, and this state is controlled by a horizontal timing circuit 58 and a vertical timing circuit 60. The differential absolute value received from the correlators 102a-102d are added together by a cumulative adder 104. Then a motion vector is produced by a motion vector generating circuit 106 and outputted. The subject is tracked based on 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 detection circuit and a subject tracking camera device using the same, and more particularly to a motion vector used for a video camera or the like which automatically follows a subject using representative point matching. The present invention relates to a detection circuit and a subject tracking camera device using the detection circuit.

[0002]

2. Description of the Related Art A conventional subject tracking camera device is
For example, JP-A-4-170870 (H04N 5/22)
5). FIG. 22 illustrates this conventional technique.
It is an illustration figure for doing. 22 in FIG. 22 (A) is a buoy
Finder's field of view, 2 is the movement displayed in a frame in field of view 1.
The detection area is shown. Further, FIG. 22B shows the motion detection area 2
Block B in which the pixels inside are divided into vertical and horizontal m × n₁₁~ B_mn
And the representative point R selected in the central pixel of each block B
₁₁~ R_mnIndicates. Furthermore, FIG.
Block B _ijRepresentative point R_ijData R_ij00And its surroundings
Pixel data S_ijxyIndicates.

Representative point R _ij of block B _ij one frame before
_Is R _ij00 , the horizontal coordinate of each pixel in the block B _ij of the current frame is x, and the vertical coordinate is y, the data of each pixel is S _ijxy , and the cumulative addition value D _xy is It is represented by the number 1.

[0004]

[Equation 1]

The xy coordinate value having the smallest D _xy is determined to be the motion vector of the subject during one frame. This motion vector indicates the direction viewed from the representative point R and the amount of motion, and the size of the block B centered on the representative point R indicates the range in which the motion of the image can be detected. As the number of cumulative addition blocks increases, the motions of the subject in the motion vector detection area 2 can be averaged, and the motion vector detection accuracy improves.

[0006]

In the above-mentioned prior art,
In order to improve the detection accuracy of the motion vector of the subject, it is necessary to increase the cumulative addition block number of the motion vector. However, if the cumulative addition block number is increased, the range of the motion vector detection area becomes wider. May be relatively small with respect to the motion vector detection area. In this case, since image data unrelated to the subject may be detected around the subject in the motion vector detection area, there is a problem that the detection accuracy deteriorates.

Therefore, a main object of the present invention is to provide a motion vector detection circuit and a subject tracking camera device using the same, which can improve the motion vector detection accuracy.

[0008]

According to a first aspect of the present invention, a plurality of detection blocks are provided in a motion vector detection area, and a motion vector of a subject is detected by using a representative point included in each detection block. It is a circuit, and is a motion vector detection circuit provided with the arrangement | positioning means which arrange | positions a detection block in piles.

A second aspect of the present invention is a motion vector detection circuit which provides a plurality of detection blocks in a motion vector detection area and detects a motion vector of a subject using color information of pixels included in each detection block. An integrating means for integrating the color information of the pixels for each detection block, and a means for detecting the motion vector of the subject based on the comparison between the color information of the specific detection block and the integrated value of the color information for each detection block, This is a motion vector detection circuit.

A third invention is a subject tracking camera device using such a motion vector detecting circuit.

[0011]

In the first aspect of the invention, the arranging means controls the latch timing of the plurality of latch means by the timing means to arrange the plurality of detection blocks in the motion vector detection area in an overlapping manner. Therefore, the number of representative points increases without expanding the range of the motion vector detection area.

According to the second aspect of the invention, the movement of the object is detected by setting a specific detection block and tracking the color information (including a plurality of color elements) detected in the detection block. At this time, a plurality of detection blocks are provided in addition to the specific detection block, and the color information is tracked based on the comparison between the color information of the specific detection block and the color information of each of the other detection blocks. It should be noted that detection blocks other than the specific detection block may be arranged so as to overlap with each other, and in this case, the number of other detection blocks to be compared with the color information of the specific detection block increases, so tracking of the color information is performed. Can be performed with higher accuracy.

[0013]

According to the first aspect of the present invention, the number of detection blocks can be increased without widening the range of the motion vector detection area. Therefore, even if the object is small, the motion vector detection accuracy can be increased, and thus the object tracking can be achieved. Can be performed with high precision. According to the second aspect, the movement of the detection block is detected by using the color information including a plurality of color elements, so that the degree of recognition of the subject is increased. Therefore, the detection accuracy of the motion vector can be increased without widening the range of the motion vector detection area, and thus the object tracking can be performed with high accuracy.

The above-mentioned objects, other objects, features and advantages of the present invention will become more apparent from the following detailed description of the embodiments with reference to the drawings.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a subject tracking camera device 10 constructed using a motion vector detection circuit 40 (see FIG. 3). The subject tracking camera device 10 includes a lens barrel 12. Figure 2
As also shown in FIG.
However, the image pickup devices 16 are mounted on the rear portions. The lens barrel 12 is supported by the first frame 22 so as to be rotatable in the yawing direction by the shaft 20. In addition, the second frame 24 uses the first frame 22 as the shaft 2.
The first frame 22 is supported so that it can be rotated in the pitching direction by 5. The second frame 24 is fixed to the housing 18, and therefore the lens barrel 12 is housed via the first frame 22 and the second frame 24 so as to be rotatable in the yawing direction and the pitching direction. It is supported by the body 18. First actuator 26
Is the coil 28 fixed to the first frame 22 and the second
Including a magnet 30 fixed to the frame 24 of
The lens barrel 12 is rotationally driven in the yawing direction with respect to the housing 18. The second actuator 32 is connected to the first frame 2
2 includes a magnet 34 fixed to the lens barrel 2 and a coil 36 fixed to the lens barrel 12, and drives the lens barrel 12 to rotate with respect to the housing 18 in the pitching direction. Then, the image signal output from the image pickup device 16 by picking up an image of the subject is the image processing circuit 3
It is converted into a video signal by 8. The motion vector detection circuit 40 detects a motion vector indicating the motion direction and size of the image in the motion vector detection area from this video signal. The motion vector detection circuit 40 will be described in detail later.

The actuator control circuit 42 drives and controls the actuators 26 and 32 in accordance with the motion vector from the motion vector detection circuit 40 to follow the subject. In addition, the zoom position detection circuit 44 provided in the lens barrel 12 includes a variable resistor, the value of the variable resistance changes according to the movement of the zoom lens 14, and a voltage corresponding to that value is given to the actuator control circuit 42. . In this way, the zoom position detection circuit 44 detects the zoom position. The motion vector detection area of the motion vector detection circuit 40 is set by the motion vector detection area control circuit 4
6 is performed. Further, the motion vector detection area superimposing circuit 50 displays the motion vector detection area on the screen of the viewfinder 48 for monitoring the image being photographed. That is, the motion vector detection area control circuit 4
6. The motion vector detection area superimposing circuit 50 and the viewfinder 48 constitute a means for setting a motion vector detection area on the screen in the viewfinder 48, and the motion vector detection area control circuit 46 displays it in the viewfinder 48. The motion vector detection area corresponding to the selected motion vector detection area is set in the motion vector detection circuit 40.

Such a subject tracking camera device 10
In the above, attention is paid to the motion vector detection circuit 40. The motion vector detection circuit 40 is configured as shown in FIG. 3 using, for example, a representative point method. Here, for example, as shown in FIG. 4, the screen 5 of the viewfinder 48 is displayed.
A case will be described in which the substantially square motion vector detection area 54 is set in 2 and the motion vector detection area 54 includes a total of 16 representative points of “◯ × Δ □”. The representative point is set at the center of each detection block, for example.

The motion vector detection circuit 40 shown in FIG.
The / D converter 56 is included. The A / D converter 56 converts the supplied video signal into digital data for each pixel. This digital data is, for example, brightness data. The video signal is also given to the horizontal timing circuit 58 and the vertical timing circuit 60. The video signal is given to the H-SYNC separation circuit 62 and the horizontal address counter 64 of the horizontal timing circuit 58. The horizontal address counter 64 counts the number of pixels in the horizontal direction according to the video signal and is reset line by line based on the horizontal synchronizing signal from the H-SYNC separation circuit 62. The output from the horizontal address counter 64 is given to the horizontal decoder 66, and the horizontal decoder 66 outputs five types of horizontal signals H0 to H5 as described later.

Further, the video signal is a vertical timing circuit 6
0 is supplied to the V-SYNC separation circuit 68 and the vertical address counter 70. The vertical address counter 70
The number of pixels in the vertical direction is counted according to the video signal, and reset for each field based on the vertical synchronizing signal from the V-SYNC separation circuit 68. Vertical address counter 7
The output from 0 is given to the vertical decoder 72, and the vertical decoder 72 outputs the vertical initial signal Vi and four kinds of vertical signals V0 to V3. Vertical initial signal Vi
When a tracking mode switch (not shown) of the subject tracking camera device 10 is turned on, is output for one field period.

The digital data of the video signal is
Based on the horizontal signals H0 to H5, the vertical signals V0 to V3, and the vertical initial signal Vi, the latch circuits 74 and 7 are provided.
Latches 6a to 76d. That is, the AND gate 78 connected to the latch circuit 74 is connected to FIG.
The horizontal signal H0 and the vertical signal V0 as shown in FIG. 4 are given, and as a result, the data of 16 representative points are latched by the latch circuit 74.

An AN connected to the latch circuit 76a
The horizontal signal H as shown in FIG.
1, the vertical initial signal Vi is applied through the vertical signal V1 and the inverter 81a. As a result, the latch circuit 76a
Thus, the data of the four detection blocks 82a each having a representative point of "O" in the motion vector detection area 54 are latched. Similarly, the AND gate 84 connected to the latch circuit 76b has a horizontal signal H2, a vertical signal V1, and an inverter 81 as shown in FIG. 5C.
A vertical initial signal Vi is given via b. As a result, the latch circuit 76b latches the data of the four detection blocks 82b each having the representative point of "x" in the motion vector detection area 54. In addition, the latch circuit 76
The vertical initial signal Vi is applied to the AND gate 86 connected to c through the horizontal signal H1, the vertical signal V2, and the inverter 81c as shown in FIG. As a result, the latch circuit 76c causes the motion vector detection area 5
The data of four detection blocks 82c each having a representative point of "Δ" in 4 are latched. Further, the AND gate 88 connected to the latch circuit 76d is connected to the AND gate 88 shown in FIG.
The vertical initial signal Vi is applied via the horizontal signal H2, the vertical signal V2, and the inverter 81d as shown in FIG. As a result, the latch circuit 76d latches the data of the four detection blocks 82d each having the representative point of "□" in the motion vector detection area 54.

It should be noted here that the detection blocks 82a to 82d can be seen with reference to FIGS.
That is, they are arranged in an overlapping manner. Here, the areas of the detection blocks 82a to 82d are set to be equal, and the degree of overlap between the detection blocks 82a and 82b is half the area of the detection block 82a (or 82b), for example. The degree of this overlap is half of the area of each detection block between the detection blocks 82b and 82d, 82a and 82c, 82c and 82d. In this way, if the degree of overlap of the detection blocks is half the area of each detection block, one pixel is included in the four detection blocks 82a and 82b at the maximum, and therefore, it corresponds to the detection blocks 82a and 82b. The required latch circuit, correlator, etc. are all needed in four sets. Therefore, in this case, many representative points can be secured without increasing the circuit scale. The degree of overlap of the detection blocks is adjusted by controlling the horizontal and vertical signals output from the horizontal timing circuit 58 and the vertical timing circuit 60, respectively.

The data of the representative points of the detection blocks 82a to 82d latched by the latch circuit 74 are:
It is stored in the representative point memory 90. Data of representative points for two consecutive fields is stored in the representative point memory 90. Therefore, the representative point data of the previous field is read in the next field, that is, the current field.

Then, the data of the read representative points is
It is latched by the latch circuits 92a to 92d.
That is, the horizontal signal H3 and the vertical signal V1 as shown in FIG. 6A are given to the AND gate 94 connected to the latch circuit 92a, and as a result, the latch circuit 92a has four representative points of "O". Latch data and correlator 10
Give to 2a. A horizontal signal H4 and a vertical signal V1 as shown in FIG. 6B are applied to the AND gate 96 connected to the latch circuit 92b, and as a result, the latch circuit 92 is provided.
b latches the data of the four representative points of "x" and supplies the data to the correlator 102b. Similarly, the AND gate 98 connected to the latch circuit 92c is supplied with the horizontal signal H3 and the vertical signal V2 as shown in FIG. 6C, and as a result, the latch circuit 92c has four representative points of “Δ”. Data is latched and given to the correlator 102c. In addition, the latch circuit 92
A horizontal signal H4 and a vertical signal V2 as shown in FIG. 6D are applied to the AND gate 100 connected to d, and as a result, the latch circuit 92d latches the data of the four representative points of "□". To the correlator 102d.

In the correlator 102a, the data of the four representative points "O" one field before, the data of the four detection blocks 82a from the latch circuit 76a, and the correlator 102b.
Then, the data of the four representative points of "X" one field before and the data of the four detection blocks 82b from the latch circuit 76b, and the data of four "Δ" one field before in the correlator 102c.
The data of one representative point and the data of four detection blocks 82c from the latch circuit 76c, the data of four representative points of "□" one field before in the correlator 102d and the data of the four detection blocks 82d from the latch circuit 76d. For each, the absolute difference value is calculated.

The correlators 102a to 102d calculate, for each detection block, the absolute difference value between the data of the representative point of the detection block and the data of each pixel in the detection block. The cumulative adder 104 adds the absolute difference values calculated for each pixel in each detection block for each pixel at the same coordinates. That is, the cumulative adder 104 has addresses for the coordinates of all pixels in the detection block, and the added value of the absolute difference values calculated for the pixels of the same coordinates in each detection block is written at each address. . Therefore, when the subject is stationary, the added value of the addresses corresponding to the representative points becomes the minimum value, and when the subject moves, the direction from the representative point to the position corresponding to the minimum value address is the direction of the motion vector. And the distance becomes the magnitude of the motion vector.

This will be described with reference to FIG. 7. The representative point data of the detection block one field before is R ₀₀ ,
_Letting S _xy be the data of each pixel in which the horizontal coordinate of each pixel in the detection block of the current field is x and the vertical coordinate is y, S _xy represents (R ₀₀ −S _xy ) for each pixel for each detection block. To calculate. This difference absolute value is added by the cumulative adder 104 for all the detection blocks to obtain the added value of each address in the cumulative adder 104. In this embodiment, as shown in FIG. 6, 16 detection blocks are included,
As shown in FIG. 7, 16 × 16 = 2 in the detection block.
It contains 56 pixels.

When the horizontal signal H5 and the vertical signal V3 as shown in FIG. 8 are applied to the AND gate 108 connected to the motion vector generation circuit 106, the motion vector generation circuit 106 causes the motion vector detection area 54 The end point is detected and the motion vector calculation is started. The motion vector generation circuit 106 obtains the motion vector of the subject in one field based on the address having the smallest addition value in the cumulative adder 104, and outputs the motion vector. This motion vector indicates the direction and the amount of motion seen from the representative point. Further, the size of the detection block centering on the representative point indicates the range in which the motion of the subject can be detected.
The movements of the subject within 4 can be averaged, and the motion vector detection accuracy is improved. The lens barrel 12 of the subject tracking camera device 10 is moved based on the motion vector.

The operation of the motion vector detecting circuit 40 thus constructed will be described with reference to FIGS. 9 and 10. In step S1 in FIG. 9, when the video signal is input, the A / D converter 56 converts the video signal into digital data for each pixel. Then, if the vertical signal V0 is output in step S3, the process proceeds to step S5. When the horizontal signal H0 is output in step S5, the process proceeds to step S7. In step S7, the representative point data of each detection block from the latch circuit 74 is stored in the representative point memory 90, and the process proceeds to step S9. When the vertical signal V0 is not output in step S3 and when the horizontal signal H0 is not output in step S5, the process proceeds to step S9.

Next, in step S9, if the field is the initial field, the process for the next pixel is performed. In step S9, the vertical decoder 72 outputs the vertical initial signal Vi (this is 1
If it is output during the field period), it is determined to be the initial field, and the latch circuits 76a to 76d are in the non-enabled state. On the other hand, if the vertical initial signal Vi is not output, it is determined that it is not the initial field, and the latch circuits 76a to 76d can be enabled based on other horizontal signals and vertical signals.

In step S9, if it is not the initial field, the process proceeds to step S11. In step S11, if it is not within the motion vector detection area 54, the processing for the next pixel is performed. In step S11, AND gates 78 and 8
The outputs from 0, 84, 86 and 88 determine that the latch circuits 74, 76a to 76d are all in the non-enabled state and are outside the motion vector detection area 54. At this time, the motion vector detection circuit 40 shown in FIG. 3 does not operate, and the process proceeds to the next pixel. On the other hand, AND gates 78 and 8
Latch circuit 7 from any of 0, 84, 86 and 88
When an enable signal for enabling the signals 4, 76a to 76d is output, the motion vector detection area 5
It judges that it is within 4 and proceeds to step S13.

When the vertical signal V1 is output in step S13, the process proceeds to step S15, and step S1
In step 5, if the horizontal signal H1 is further output, the process proceeds to step S17. In step S17, the correlation absolute value is calculated by the correlator 102a from the data of the representative point "O", and then the process proceeds to step S19. In step S15, the process also goes to step S19 when the horizontal signal H1 is not output. In step S19, the horizontal signal H2
If is output, in step S21, the correlator 102b calculates the absolute value of the difference from the data of the representative point of "x", and the process proceeds to step S23 shown in FIG. Step S
When the vertical signal V1 is not output at 13 and the horizontal signal H2 is not output at step S19, the process proceeds to step S23.

In step S23, if the vertical signal V2 is output, the process proceeds to step S25 and step S2.
If the horizontal signal H1 is output at 5, the absolute value of the difference from the data of the representative point of “Δ” is calculated in the correlator 102c at step S27, and step S2
Proceed to 9. When the horizontal signal H1 is not output in step S25, the process also goes to step S29. If the horizontal signal H2 is output in step S29, the absolute value of the difference from the data of the representative point of "□" is calculated by the correlator 102d in step S31, and step S33.
Proceed to. Even when the vertical signal V2 is not output in step S23 and the horizontal signal H2 is not output in step S29, the process proceeds to step S33.

In step S33, the cumulative adder 10
In step 4, the difference absolute value is added for each same address of each detection block, and the process proceeds to step S35. In step S35,
If both the horizontal signal H5 and the vertical signal V3 are applied to the AND gate 108 connected to the motion vector generation circuit 106, it is determined to be the end of the motion vector detection area, and the process proceeds to step S37. In step S37, the motion vector generation circuit 106 generates a motion vector, the motion vector is output in step S39, and the process proceeds to the next pixel. Even when the end of the motion vector detection area 54 is not detected in step S35, the process proceeds to the next pixel.

In the subject tracking camera device 10 including such a motion vector detecting circuit 40, since the detection blocks can be arranged in an overlapping manner, the representative point interval can be narrowed regardless of the size of the detection blocks. . Therefore, the number of representative points in the motion vector detection area 54 increases significantly. As a result, unlike the conventional technique in which the detection blocks are arranged next to each other, the detection block interval and the representative point interval do not match, so that even if the number of representative points is increased, the motion vector detection area 54 is enlarged by the detection block interval. There is no such thing.

In addition, in the subject tracking camera device 10, as shown in FIG.
You may use the motion vector detection circuit 110 as shown in FIG. The motion vector detection circuit 110 shown in FIG. 11 is configured using the representative color method and does not overlap the detection blocks. Here, as shown in FIG. 12, the motion vector detection area 112 is divided into a plurality (25 in this embodiment) of detection blocks 114, and the motion vector is calculated based on the color information of the pixel of the center block 114a. The case of obtaining will be described. In addition, in the color information, here,
It includes three color elements of Y, RY, and BY. Also,
Horizontal decoder 6 included in motion vector detection circuit 110
6 outputs horizontal signals H0 to H3, which will be described later, and the vertical decoder 72 outputs a vertical initial signal Vi and vertical signals V0 and V2, which will be described later.

A video signal is input to the motion vector detection circuit 110, converted into digital data by the A / D converter 56, and then digital data of each detection block 114 in the motion vector detection area 112 is converted by the latch circuit 116. To latch. In the initial field where the horizontal signal H0 and the vertical signal V0 are output as shown in FIG. 12, the enable signal is applied to the latch circuit 116 through the AND gate 118 and the OR gate 120. Then, the latch circuit 116 latches the data of each pixel of the detection block 114a and supplies it to the block integration circuit 122.

The operation of the block integration circuit 122 is shown in FIG.
Will be described with reference to. The detection block 114 illustrated in FIG. 13A includes 8 × 8 = 64 pixels. The respective values of the three color elements constituting the color information of each pixel, y _ij, as shown in FIG. 13 (A), (r- y) ij, (b-
y) _ij , the data of the representative color is expressed as in Equation 2.

[0039]

[Equation 2]

This representative color data is stored in the representative color memory 124.
Stored in. The representative color memory 124 stores representative field data of one initial field, and the data of the initial field is output in the subsequent fields, that is, the current field. Therefore, the correlator 136 described later calculates the representative color data of the previous field and the integrated value data of each detection block 114 of the current field.

On the other hand, in the fields other than the initial field, the vertical initial signal Vi is not output, so the inverter 1
The inverted output via 26 is provided to AND gate 128. Further, when the horizontal signal H1 and the vertical signal V1 are applied to the AND gate 128, the enable signal is applied to the latch circuit 116 via the OR gate 120, and the latch circuit 116 is enabled. Then, the latch circuit 116 latches the data of the pixels of all the 25 detection blocks 114 and supplies them to the block integration circuit 122. In the block integration circuit 122, color information is integrated for each of the 25 detection blocks 114 shown in FIG. The integrated value data obtained by integrating the color information of the pixels for each of the detection blocks 114 is represented by Expression 3.

[0042]

[Equation 3]

In the integrated value memory 130, the detection block 1
The integrated value data for 14 pieces (25 pieces in this embodiment) is stored. When the horizontal signal H2 and the vertical signal V2 as shown in FIG. 14 are applied to the AND gate 134 connected to the latch circuit 132, the representative color data in the representative color memory 124 is supplied to the correlator 136. When the horizontal signal H3 and the vertical signal V2 as shown in FIG. 15 are applied to the AND gate 140 connected to the latch circuit 138, the integrated value data for each detection block in the integrated value memory 130 is latched, Correlator 13
Given to 6. The horizontal signal H3 is output for the number of detected blocks, that is, 25 horizontal signals. The correlator 136 calculates the representative color data of the initial field and the integrated value data of each detection block of the current field. That is, the sum D _{xy of the} absolute difference values for each color element is calculated by the equation 4.

[0044]

## EQU4 ## D _xy = | R _xy −R0 | + | (R−Y) _xy − (R
-Y) 0 | + | (BY) _xy- (BY) 0 | This is detected by each detection block 11 as shown in FIG. 13 (B).
The motion vector generating circuit 142 compares the respective calculation results, determines that the center (114a) of the detection block 114 has moved to the position where D _xy has the minimum value, and obtains the motion vector. . Then, the motion vector generation circuit 142 outputs the motion vector. as a result,
In the next field, the lens barrel 12 of the subject tracking camera device 10 is moved based on the motion vector.

The operation of the representative color type motion vector detection circuit 110 will be described with reference to FIG. Figure 1
In step S51 shown in FIG. 6, if the tracking mode switch of the subject tracking camera device 10 is turned on and the vertical initial signal Vi is output, it is determined to be the initial field, as in the above-described embodiment. At this time, the enable signal is not output from the AND gate 128.

If it is the initial field in step S51, the process proceeds to step S53. If the vertical signal V0 is output in step S53, the process proceeds to step S55. If the horizontal signal H0 is output in step S55, the representative color data is integrated by the block integrating circuit 122 in step S57, and the data is stored in the representative color memory 124 in step S59.
Proceed to processing the next pixel. If the vertical signal V0 is not output in step S53 and the horizontal signal H0 is not output in step S55, the process proceeds to the next pixel.

On the other hand, in step S51, if it is not the initial field, the process proceeds to step S61. Step S6
If the vertical signal V1 is output at 1, the process proceeds to step S63. In step S63, the horizontal signal H1
Is output, in step S65, the color information of each detection block 114 is integrated by the block integration circuit 122, and the integrated value data is stored in the integrated value memory 130. Next, in step S67, it is determined whether or not it is the end of the motion vector detection area 112. This is determined by whether or not the enable signal is input to the latch circuits 132 and 138. If the enable signal is input, the correlator 13 is output from the latch circuits 132 and 138.
The representative color data of the representative color memory 124 and the integrated value data of each detection block 114 are given to 6, respectively, and the absolute difference value is calculated by the correlator 136 in step S69.

Next, in step S71, the motion vector generation circuit 142 generates a motion vector, and in step S73 the motion vector is output, and the process proceeds to the next pixel. Further, in step S61, the vertical signal V
When the horizontal signal H1 is not output in step S63, or when the end of the motion vector detection area 112 is not detected in step S67, the process proceeds to the next pixel.

In the subject tracking camera device 10 using the motion vector detection circuit 110, the movement of the subject is detected as the movement of the detection block 114a in the motion vector detection area 112, and the lens barrel 12 is tracked by the detection block 114a. The direction of (camera) was controlled. Therefore, even when a plurality of objects having the same color as the target exist in the screen 52, if only one object exists in the motion vector detection area 112, only the target object can be tracked well.

Further, a motion vector detecting circuit 150 as shown in FIG. 17 may be used in the subject tracking camera device 10. This motion vector detection circuit 150 is of a representative color system and is a case where detection blocks are overlapped. Here, as shown in FIG. 18, a detection block 152 indicated by a broken line is superimposed on the detection block 114 similar to that of the embodiment of FIG.
Are arranged. There are 25 detection blocks 114 and 16 detection blocks 152. Each detection block 114
And 152 have the same area, and the detection block 1
Here, the degree of overlap between 14 and the detection block 152 is about 1/4 of the area of the detection block 114 (detection block 152). When the detection blocks are overlapped as shown in FIG. 18, the distance between the central detection block 114a and each of the upper, lower, left, and right detection blocks 152 becomes uniform, and the detection block 114a and the diagonally upper left, diagonally upper right, diagonally lower left,
The distance from each of the detection blocks 152 on the lower right is uniform. That is, since the detection blocks 152 can be arranged vertically and horizontally symmetrically with respect to the detection block 114a, the motion vector can be detected with the same accuracy in any direction.

This motion vector detection circuit 150 is shown in FIG.
In the motion vector detection circuit 110 shown in FIG.
4, a block integration circuit 156, an AND gate 158 and an inverter 160 are added. The other configurations are similar to those of the motion vector detection circuit 110, and thus their duplicated description will be omitted. In this motion vector detection circuit 150, when the latch circuit 154 is enabled, each detection block 1 from the A / D converter 56 is detected.
The digital data of 52 is latched, and the block integration circuit 156 performs the same calculation as that of the block integration circuit 122 described above. The integrated value data, which is the calculation result in the block integrating circuit 156, is stored in the integrated value memory 130. That is, the AND gate 15 connected to the latch circuit 154 in a state where the vertical initial signal Vi is not output.
18, when the horizontal signal H2 and the vertical signal V2 as shown in FIG. 18 are applied, the latch circuit 154 is enabled and the data of the detection block 152 shown by the broken line in FIG. 18 is latched. The horizontal decoder 66 outputs horizontal signals H0 to H4, and the vertical decoder 72 outputs a vertical initial signal Vi and vertical signals V0 to V3. Further, the AND gate 134 connected to the latch circuit 132 is supplied with the horizontal signal H3 and the vertical signal V3, respectively, as shown in FIG.
A horizontal signal H4 and a vertical signal V3 are applied to the AND gate 140, which is connected to, as shown in FIG. The horizontal signal H4 is output by the number of detection blocks (25 + 16 = 41 in this embodiment).

The operation of the motion vector detecting circuit 150 will be described with reference to FIG. In step S101 shown in FIG. 21, similarly to the previous embodiment, it is determined whether the field is the initial field or not based on the vertical initial signal Vi. In the case of the initial field, the vertical decoder 72 outputs the vertical initial signal V
Since i is output, AND gates 128 and 1
No enable signal is output from 58. Step S
In 101, if it is the initial field, step S1
Go to 03. In step S103, the vertical signal V0
If is output, the process proceeds to step S105. If the horizontal signal H0 is output in step S105, the process proceeds to step S107. That is, the latch circuit 116 is enabled and integrated by the block integration circuit 122. In step S109, the representative color data is stored in the representative color memory 124, and the process for the next pixel is performed. In step S103, the vertical signal V0 is changed to step S1.
In 05, even when the horizontal signal H0 is not output, the process proceeds to the processing of the next pixel.

On the other hand, in step S101, if the vertical initial signal Vi is not output and it is not the initial field, the process proceeds to step S111. In step S111, if the vertical signal V1 is output, step S113
Proceed to. When the horizontal signal H1 is output in step S113, the process proceeds to step S115. That is, the latch circuit 116 is enabled, and the block integration circuit 122 operates. This is a calculation for each of the 25 detection blocks 114 shown by the solid line in FIG. The integrated value data here is stored in the integrated value memory 130, respectively.

Next, in step S117, if the vertical signal V2 is output, the process proceeds to step S119.
In step S119, if the horizontal signal H2 is output, the process proceeds to step S121. That is, the latch circuit 154 is enabled and is calculated by the block integration circuit 156 in step S121. this is,
The calculation is performed for each of the 16 detection blocks 152 indicated by the broken line in FIG. The integrated value data here is stored in the integrated value memory 130, respectively. Then, the process proceeds to step S123.

In step S123, it is determined whether it is the end of the motion vector detection area 112. This is determined by whether or not the enable signals are applied to the latch circuits 132 and 138, respectively. When these enable signals are given and the end of the motion vector detection area 112 is detected, in step S125, the correlator 13
The absolute difference value is calculated at 6. Then, step S12
In 7, the motion vector generation circuit 142 generates a motion vector, the motion vector is output in step S129, and the process proceeds to the next pixel. Even when the vertical signal V1 is not output in step S111 and the horizontal signal H1 is not output in step S113, or when the end of the motion vector detection area 112 is not detected in step S123,
Proceed to processing the next pixel.

By superposing the detection block 152 on the detection block 114, the motion vector can be detected with higher accuracy. That is, the motion vector detection circuit 110 shown in FIG. 11 detects the motion vector in the detection block unit shown in FIG. 12, but the motion vector detection circuit 150 can detect the motion vector finer and with higher accuracy. .

Using the representative color method as described above, FIG.
Since the motion vector detection circuit 110 shown in FIG. 11 and the motion vector detection circuit 150 shown in FIG. 17 use color information including a plurality of color elements, the object is more preferable than the representative point method that uses only single information such as brightness. Is highly recognizable. Further, since the motion vector detection area 112 is set, the degree of misidentifying an object other than the subject as the subject is low. Furthermore, since several tens to several hundreds of pixels are set as one detection block and the integrated value data of the color information for each detection block is used, the calculation amount and the memory amount can be relatively small.

In the above embodiment, the color information is
Three color elements of Y, RY and BY were used. This is effective when processing a video signal or the like. In addition to the above, when processing a signal from a primary color CCD or the like, R,
When processing a CCD or the like in which a mosaic filter is arranged in a complementary color system using three color elements G and B, Y and Cr are used.
It is preferable to use three color elements (= R-2G) and Cb (= B-2G).

Further, R, G and B are respectively divided by Y,
Three color elements of R / Y, G / Y and B / Y may be used. By thus dividing by Y and normalizing, it is possible to remove the influence of the lightness due to the difference in illumination, etc., and if the hue and the saturation are the same, they can be regarded as the same color. The motion vector detection circuit 40 shown in FIG.
In the motion vector detection circuit 150 shown in FIG. 1, it is sufficient if the detection blocks overlap in at least one of the horizontal direction and the vertical direction.

Further, in the above-mentioned embodiment, the case where the motion vector detecting circuit is applied to the subject tracking camera device 10 has been described. However, the motion vector detecting circuit of the present invention can be applied to a video camera having an image stabilizing device. Needless to say.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a subject tracking camera device according to an embodiment of the present invention.

FIG. 2 is a vertical cross-sectional view showing an imaging portion of FIG.

FIG. 3 is a block diagram showing an example of a motion vector detection circuit.

FIG. 4 is an illustrative view showing a motion vector detection area, a representative point, a horizontal signal and a vertical signal used in the motion vector detection circuit of FIG.

5 is an illustrative view showing a data input timing of the latch circuit of FIG. 3. FIG.

FIG. 6 is an illustrative view showing an input timing of representative point data of the other latch circuit of FIG.

FIG. 7 is an illustrative view showing a detection block and a pixel.

FIG. 8 is an illustrative view showing a motion vector calculation start timing.

9 is a flowchart showing an example of the operation of the motion vector detection circuit of FIG.

FIG. 10 is a flowchart showing a continuation of the operation of FIG.

FIG. 11 is a block diagram showing another motion vector detection circuit of the present invention.

12 is an illustrative view showing a motion vector detection area, a detection block, a horizontal signal, and a vertical signal used in the motion vector detection circuit of FIG.

13A is an illustrative view showing a detection block and pixels in FIG. 12, and FIG. 13B is an illustrative view showing one example of a motion vector.

FIG. 14 is an illustrative view showing an input timing of representative color data of the latch circuit of FIG.

FIG. 15 is an illustrative view showing an input timing of integrated value data of the other latch circuit of FIG. 11.

16 is a flowchart showing one example of the operation of the embodiment in FIG.

FIG. 17 is a block diagram showing another motion vector detection circuit of the present invention.

18 is an illustrative view showing a motion vector detection region, a detection block, a horizontal signal and a vertical signal used in the motion vector detection circuit of FIG.

19 is an illustrative view showing an input timing of representative color data of the latch circuit of FIG.

20 is an illustrative view showing an input timing of integrated value data of the other latch circuit of FIG.

FIG. 21 is a flowchart showing an example of the operation of the embodiment in FIG.

FIG. 22 is an illustrative view for explaining a conventional technique.

[Explanation of symbols]

10 ... Subject tracking camera device 14 ... Zoom lens 16 ... Image sensor 38 ... Image processing circuit 40 ... Motion vector detection circuit 42 ... Actuator control circuit 46 ... Motion vector detection area control circuit 48 ... Viewfinder 50 ... Motion vector detection area superposition circuit 58 ... Horizontal timing circuit 60 ... Vertical timing circuit 74, 76a to 76d, 92a to 92d, 116, 13
2, 138, 154 ... Latch circuit 90 ... Representative point memory 102a-102d, 136 ... Correlator 104 ... Cumulative adder 106, 142 ... Motion vector generation circuit 124 ... Representative color memory 130 ... Integrated value memory

Claims (13)

[Claims] 1. A motion vector detection circuit, wherein a plurality of detection blocks are provided in a motion vector detection area, and a motion vector of a subject is detected by using a representative point included in each detection block. Equipped with an arrangement means for arranging them in an overlapping manner,
Motion vector detection circuit. 2. The arranging means comprises a plurality of latch means for holding video signals of different detection blocks, and the latch timings so that the detection blocks of the video signals held by the latch means overlap each other. The motion vector detection circuit according to claim 1, comprising timing means for controlling the means. 3. The motion vector detection circuit according to claim 1, wherein the plurality of detection blocks are arranged to overlap each other in at least one of a horizontal direction and a vertical direction. 4. The overlap between the detection block and an adjacent detection block is about half the area of the detection block.
The motion vector detection circuit according to claim 1. 5. A motion vector detection circuit, wherein a plurality of detection blocks are provided in a motion vector detection area, and a motion vector of a subject is detected by using color information of pixels included in each detection block. An integrating unit that integrates the color information of pixels for each block, and a unit that detects the motion vector of the subject based on a comparison between the color information of a specific detection block and the integrated value of the color information of each detection block are provided. Motion vector detection circuit. 6. The motion vector detection circuit according to claim 5, wherein the color information includes three types of different color elements. 7. The three types of color elements are Y, RY, B-
7. The motion vector detection circuit according to claim 6, which is Y. 8. The motion vector detection circuit according to claim 6, wherein the three types of color elements are Y, Cr, and Cb. 9. The motion vector detection circuit according to claim 5, further comprising an arrangement unit that arranges the plurality of detection blocks in the motion vector detection area so as to overlap each other. 10. The arranging means comprises a plurality of latch means for holding video signals for different detection blocks, and the latch timings so that the detection blocks of the video signals held by the respective latch means overlap each other. 10. The motion vector detection circuit according to claim 9, comprising timing means for controlling the means. 11. The motion vector detection circuit according to claim 9, wherein the plurality of detection blocks are arranged to overlap each other in at least one of a horizontal direction and a vertical direction. 12. The motion vector detection circuit according to claim 9, wherein an overlap between the detection block and an adjacent detection block is about ¼ of the area of the detection block. 13. An object tracking camera device using the motion vector detection circuit according to claim 1. Description: