JPH06217187A - Image pickup device, moving vector detector, and picture shake correcting device - Google Patents

Abstract

PURPOSE:To obtain a stable reproduced picture even at the time of photographing a moving subject by recognizing the head part of a human body by a picture signal, and adjusting a zoom so that the size of a face part can be constant. CONSTITUTION:A video signal V outputted from a camera signal processing circuit 3 is inputted through an A/D converter 13 to a face picture recognizing circuit 14. The circuit 14 specifies the area of the face picture from the color, area, and shape or the like. A zoom control circuit 15 controls a zoom lens control circuit 9 and an electronic zoom circuit 16 by the information of the size and position of the face part in the picture obtained by the circuit 14. Then, when the movement of the subject is strong, the magnification of an optical zoom is decreased, and the magnification of an electronic zoom is increased. Also, when the movement of the object is small, the magnification of the optical zoom is increased, and the magnification of the electronic zoom is decreased. Thus, the moving subject can be followed-up, the deterioration of the picture quality of the electronic zoom can be reduced, and the stable picture can be obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image pickup device used in a camera-integrated VTR or the like, a motion vector detection device for detecting a motion of an image from a video signal, and a shake of the entire image due to vibration of a video camera or the like. The present invention relates to an image shake correction device for correction.

[0002]

2. Description of the Related Art In recent years,
Automatic video movie control technology has become indispensable. Currently, some single-lens reflex cameras have an auto zoom function that automatically adjusts the zoom according to the distance to the subject. The auto zoom function used in the single-lens reflex camera will be described below.

FIG. 38 is a block diagram showing a conventional image pickup apparatus. In the figure, reference numeral 1 denotes a lens barrel, which includes a zoom lens 21 and a focus lens 22. In reality, it is composed of more lenses, but these lenses are omitted. The motor driver 10 drives the zoom motor 11 that moves the zoom lens 21 based on the zoom command from the zoom lens control circuit 9. Motor driver 6
Drives the motor 7 that moves the focus lens 22 based on the focus command from the focus lens control circuit 5. The zoom lens position detection device 12 is a zoom lens.
The position of 21 is detected and the signal corresponding to the position of the lens is output one-to-one. The focus lens position detection device 8 detects the position of the focus lens 22 and outputs a signal corresponding to the lens position on a one-to-one basis. A subject image is formed on the light receiving surface 2 by the zoom lens 21 and the focus lens 22. In this conventional example, a film is placed on the light receiving surface 2.

The infrared distance measuring device 23 measures the distance from the camera to the subject by emitting infrared light toward the subject and receiving the infrared light reflected by the subject. The distance information obtained by the infrared distance measuring device 23 is sent to the focus lens control circuit 5. The focus lens control circuit 5 controls the motor driver 6 based on the distance information to drive the focus lens 22. The distance information obtained by the infrared distance measuring device 23 is also sent to the zoom lens control circuit 9. The zoom lens control circuit 9 controls the motor driver 10 so that the object image on the light receiving surface 2 has a constant size based on the distance information obtained from the infrared distance measuring device 23 and the focal length information of the lens obtained in advance. The zoom lens 21 is controlled and driven. By the above operation, it is possible to always obtain a subject image of the same size on the light receiving surface 2.

The auto-zoom described above is a function originally developed for shooting a still image such as a still camera or a single-lens reflex camera, and does not consider taking a moving image. For example, when the subject moves and shifts from the center position to the edge of the screen, the infrared range finder 23 measures the distance to the background, not the distance to the subject, and the auto zoom does not operate properly. Also, if the zoom lens is on the tele side, moving the camera a little will cause the subject in the screen to move greatly, and it is not easy to follow the moving subject, and it is very difficult to make the auto zoom function sufficiently. There was a problem that it was difficult.

A conventional image stabilization mechanism used in a portable video camera or the like is, for example, Japanese Patent Laid-Open No. 61-255173.
There is one such as shown in Japanese Patent Publication. FIG. 39 is a block diagram of a conventional imaging device. In FIG. 39, 51 is a lens barrel (optical system), 52 is a gimbal mechanism that supports the lens barrel 51, and 53 is the lens barrel 51.
, 54 is an angle sensor that detects the relative angle between the lens barrel 51 and the housing (not shown), 55 is an angular velocity sensor that detects the angular velocity generated in the lens barrel 51, and 56 is the output of the angle sensor 54 A variable amplification circuit for applying a predetermined gain to 57, a variable amplification circuit for applying a predetermined gain to the output of the angular velocity sensor 55, an adder 59 for adding the outputs of the variable amplification circuits 56 and 57,
Reference numeral 60 is an actuator drive circuit for driving the actuator 53 from the output of the adder 59, 61 is an image pickup device for picking up an image of a subject through the lens barrel 51, and 62 is a signal processing circuit for processing a video signal output from the image pickup device 61.

The lens barrel 51 and the image pickup device 61 include a gimbal mechanism 52.
Is rotatably supported by two orthogonal axes. The actuators attached to the respective rotation axes control the lens barrel 51 so that it is stationary with respect to the absolute coordinate system. The sensor 55 detects the angular velocity generated in the lens barrel 51 due to disturbance vibration such as camera shake, and the actuator 53 is driven by a control value corresponding to this value. This is called an angular velocity loop. Basically, this control system realizes the image stabilization function.

On the other hand, in order to construct a practical image pickup apparatus, it is desirable that the central axis of the lens barrel 51 and the central axis of the housing be aligned. Therefore, the relative angle between the center axis of the lens barrel 51 and the center axis of the housing is detected by an angle sensor 54 such as a Hall element, and the actuator is controlled by a control value corresponding to this value. This control system is called an angle loop and operates in a frequency band lower than that of the angular velocity loop. This control loop causes the lens barrel 51 to match the central axis of the housing in the low frequency region.

In order to perform camera shake correction with such a control system, the gain of the angular velocity loop must be made relatively large as compared with the gain of the angle loop. Therefore, when the relative angle between the lens barrel and the housing is small, the output to the actuator becomes very small, and when the loss component of the rotation axis is considered, the lens barrel cannot be completely returned to the central axis. Occurs. With respect to this problem, especially when the mechanical system is small and lightweight, the loss component due to friction of the rotating shaft relatively increases, which is a serious drawback.

In recent years, there is a strong demand for downsizing and weight saving of video cameras, and for reasons such as hardware scale reduction and processing flexibility, a control circuit such as this system is also required to perform digital processing using a microcomputer. There is more to do. Therefore, if the control system is configured by a digital circuit, the number of effective bits is limited, so that a minute value of 1 bit or less is not output. In this case as well,
There is a problem that the return to the origin of the lens barrel becomes insufficient.

As a measure against such a problem, there is a method of adding an integral element to the angle loop of the control system. However, this method has a drawback that the overshoot generated at the output becomes large. This is an operation of passing the origin position when the lens barrel returns to the origin, which significantly deteriorates the operability of the image pickup apparatus. This operation is a serious drawback especially when shifting to still photography.

Conventionally, in order to compensate for motion compensation for improving coding efficiency and to correct screen shake due to camera vibration or shake, detection of a screen parallel movement amount, that is, a motion vector has been used.

However, since the actual movement of the image is a mixture of the parallel movement of the entire image and the movement of the object in the image,
The motion of the object may cause false detection of the motion vector,
There is a problem depending on the image, such as a decrease in detection accuracy due to the image pattern.

In particular, the motion vector is detected by performing a correlation operation on image information between two temporally consecutive screens, and the shift amount having the highest correlation among the correlation values obtained thereby is used as the motion vector. Therefore, in the motion vector detection when there is no change in the image or when the same pattern is repeated periodically, the possibility of erroneous detection increases.

Further, when a moving object enters the screen for the purpose of correcting the shake caused by camera shake, only the motion vector detected from the background portion should be used, and the motion vector detected from the moving object portion is used. Should not be. Therefore, it is necessary not to use the detected motion vector as it is, but to select it according to the situation.

Therefore, for example, as disclosed in Japanese Patent Laid-Open No. 61-269375, a candidate vector is calculated for each divided screen by dividing the image. For example, for a divided screen in which the pattern changes little, the candidate vector is calculated. Is determined to be invalid,
It is conventional to exclude from the final motion vector detection calculation.

FIG. 40 is a block diagram showing the structure of this conventional motion vector detecting device. In this conventional example, an example including five fixed motion detection areas (HL, HR, LL, LR, CT) will be described. In FIG. 40, 100 is a video signal input terminal, 101 is a representative point memory that stores the value of the representative point of the previous screen, and 102 is the absolute difference between the representative point of the previous screen and the pixels around the representative point of the current screen. Absolute value calculation circuit for calculating the value, 103
Is an HL area gate circuit that passes the absolute value calculation result only at the detection timing of the HL area, 104 is an HL area cumulative addition circuit for detecting the movement of the HL area, and 105 is the reliability of the detection result from the state of the HL area cumulative addition. This is an HL area reliability determination circuit for determining.

Numerals 106 to 117 are gate circuits for the HR area, LL area, LR area, and CT area, a cumulative addition circuit, and a reliability determination circuit. Reference numeral 118 is a motion vector determination circuit that determines the final motion of the entire screen from the motion vector value of each area and the reliability determination result.

In the conventional motion vector detecting apparatus shown in FIG. 40, an input terminal 100 is supplied with a temporally continuous video signal of at least two fields. As shown in 141, 142, 143, 144, 145 shown in Fig. 41, there are five motion vector detection areas (HL, HR, L respectively) in the screen.
L, LR, and CT) are determined, and the motion vector is detected for each divided screen unit by the correlation between consecutive screens.

The correlation is, for example, the representative point Rij of the previous screen.
And the absolute value of the difference between the signal Sij (x, y) of the current screen, which has a positional relationship in the horizontal direction x and vertical direction y, is obtained, and x and y having the same positional relationship for each representative point are shown below. It is obtained from Dxy added for a predetermined number of blocks. Dxy = Σ | Rij-Sij (x, y) |

Generally, the displacement (x, y) that gives the minimum value of Dxy is detected, and this is used as the motion vector. Further, the minimum value, the maximum value, and the average value of Dxy are obtained, and together with the value of the motion vector, the motion vector reliability determination circuits 105, 108,
It is input to 111, 114, 117. Dxy shows a typical pattern depending on the nature of the screen. For easy understanding,
A one-dimensional view of the relationship between the displacement (x, y) and the cumulative addition value Dxy is as shown in FIGS. 42 (a) to 42 (d).

FIG. 42 (a) shows an example of maximum value> average value >> minimum value. This shows that the reliability of the obtained motion vector value is very high in the case of an ideal design in which the design of the image changes as a whole.

In the case of FIG. 42 (b), an example of maximum value≈average value≈minimum value≈0 is shown. This is because the reliability of the obtained motion vector value is low and should be invalidated when there is no change in the pattern of the image, for example, when a blue sky or a plain wall is photographed.

FIG. 42 (c) shows an example of maximum value >> average value≈minimum value. This is a case where the change in luminance is regularly repeated, for example, when a blind is photographed, and the reliability of the motion vector value obtained in this case is low.

The case of FIG. 42 (d) shows an example of maximum value≈average value≈minimum value >> 0. This is the case where the change in the image is too large, for example, when the image is swung around the camera, and the reliability of the motion vector value obtained in this case is low.

From the above properties, the reliability judgment circuit 10
5, 108, 111, 114, 117 is the maximum value in each split screen unit,
The reliability of the motion vector is determined based on the average value and the minimum value, and the motion vector determination circuit 118 determines the average value of the motion vectors of the areas determined to have high reliability as the motion vector of the entire screen. .

In the above description, the reliability is judged from the maximum value, the average value and the minimum value of the correlation values, and the average value of the motion vectors with high reliability is determined as the motion vector of the entire screen. The reliability is obtained from the second smallest value or the state of these temporal changes, and even when it is determined that the reliability is high, the direction of each vector is checked to determine whether the subject is the same or different, If the determinations are different, various algorithms are conceivable, such as setting the motion vector of the entire screen to 0, obtaining the average, determining which is the main subject, and determining the value of the area of the main subject.

However, in the above-mentioned configuration, the area determined to have low reliability is ignored, so that
For example, when the reliability of the entire area is low, the motion vector detection value of the entire screen becomes 0. Further, when there are few highly reliable regions, there is a problem that the accuracy in obtaining the average value as the final motion vector detection value becomes poor.

In recent years, video cameras have rapidly become popular along with the progress of miniaturization and automation, and therefore, there are many opportunities for beginners who have little experience in camera shooting to use them. However, when a beginner who does not have sufficient shooting technology shoots, the captured image is often difficult to see due to image distortion due to camera shake, which is a serious problem with higher zoom lens magnification. It's starting. As a countermeasure against this, imaging apparatuses having various image stabilization functions have been conventionally proposed. An example is shown below.

FIG. 43 is a block diagram showing a conventional image shake correction apparatus. In FIG. 43, 301 is a video signal input terminal, 302 is an A / D converter for converting the video signal input to the input terminal 301 into a digital signal, and 303 is an A / D converter.
A field memory for storing one field of the video signal converted into a digital signal by 302, 304 is an A / D converter
A motion vector detection circuit that detects a motion vector for each field of the entire image from the digitally converted video signal output by 302, and 305 integrates the motion vector detected by the motion vector detection circuit 304 and reads out the address of the field memory 303. , 306 is a field memory according to the output of the address generation circuit 305.
A memory control circuit that controls the reading of 303, a 307 is an interpolation enlargement circuit that interpolates and enlarges the video signal read from the memory 303, and a 308 is a D / that converts the output signal of the interpolation enlargement circuit 307 into an analog signal and outputs it. A converter.

Next, the operation will be described. Input terminal 30
The video signal input to 1 is converted into a digital signal by the A / D converter 302. The output of the A / D converter 302 is sent to the field memory 303 and the motion vector detection circuit 304, respectively. The motion vector detection circuit 304 detects the motion vector of the image for each field by the known representative point matching method, and outputs it to the address generation circuit 305. The address generation circuit 305 obtains the read address of the field memory 303 from the motion vector of the image for each field output from the motion vector detection circuit 304, and outputs it to the memory control circuit 306. A / D in the field memory 303
One field of the video signal output by the converter 302 is written. The memory control circuit 306 controls writing / reading of the field memory 303. Here, when writing the output of the A / D converter 302 to the field memory 303,
Although the entire screen for one field is written, when reading this, a screen area smaller than the written screen is read, and the read position is changed according to the address output by the motion vector detection circuit 304. As a result, image blur can be reduced.

The principle will be described with reference to FIG. In FIG. 44, 311 is a screen written in the field memory 303, 312a and 312b are screens read from the field memory 303, and 313a and 313b are subjects in the screen. In one field, the subject 313a in the writing screen 311
However, if it moves to the position of 313b in the next field due to image shake and the motion vector at that time is Δv,
By moving the read screen from the position of 312a to the position of 312b by Δv, the subject in the read screen can be fixed at a fixed position. By repeating the above operation for each field, continuous image blur included in the captured image can be removed. Therefore, the read address of the field memory 303 output from the address generation circuit 305 may be the initial value of which the integrated value ΣΔv of the motion vector from the start of the image blur correction operation is added.

The image signal whose image blur has been corrected as described above is interpolated / enlarged to the size of the input screen in the interpolation / enlargement circuit 307, converted into an analog signal by the D / A converter 308, and output.

At present, in the field of consumer use, the type of image shake correcting apparatus is mainly of a type that is incorporated in a camera-integrated VTR to correct the image shake during photographing. One example is shown in FIG. In FIG. 45, 321 is a lens unit that forms an optical image of a subject, 322 is a photoelectric conversion unit that converts the optical image formed by the lens unit 321 into an electric signal, and 323 is a structure similar to that shown in FIG. An image shake correction unit that holds and corrects the shake of the entire image formed by the video signal output by the photoelectric conversion unit 322, and 324 is a video signal in which the shake of the image output by the image shake correction unit 323 is corrected. Recording / playback VT
It is the R part.

However, since the image shake correction apparatus shown in FIG. 43 cuts out a part of the image output from the camera unit and restores it to the original screen size by interpolation, the image quality is essentially deteriorated. Accompany. Therefore, it is desirable to use the image shake correction function only when the image shake is severe. Therefore, it is advantageous to select a portion that corrects the image shake while watching the playback screen after the image is recorded, rather than correcting the image shake when the image is recorded as in the example shown in FIG.

By the way, the image shake on the monitor screen may be caused by the shake of the photographing apparatus due to camera shake or the like, as well as the time base fluctuation of the video signal such as the reproduction jitter of the VTR. VTR for playing video is a VT for home use such as Betamax standard, VHS standard, 8 mm standard
If it belongs to the R and U-matic standards, the reproduced video signal is not subjected to the time-axis correction of the luminance signal except for some models, so that the time-axis fluctuation with respect to the regular video signal standard may be included. Many. Due to the nature of video images taken by video cameras, editing work is often performed.However, when dubbing is performed with the time axis fluctuation left untouched, the time axis fluctuations accumulate and the playback image suffers unsightly shaking due to time axis errors. There was a problem that it occurred.

The present invention has been made in view of the above circumstances, and one object of the present invention is to capture a moving image, and the subject does not necessarily have to be in the center.
An object of the present invention is to provide an imaging device that functions accurately even when moved.

Another object of the present invention is to provide an image pickup apparatus capable of achieving both high image stabilization performance and highly accurate home position return even with a control circuit having a low resolution.

Still another object of the present invention is to provide a motion vector detecting device which can detect a motion vector with high accuracy and without discomfort even in the case of a pattern having a small area where a highly reliable motion vector can be detected, and this device. An object is to provide an image shake correction device using the same.

Still another object of the present invention is to provide an image shake correction apparatus capable of correcting the fluctuation of the time axis and the shake of the entire image.

[0041]

An image pickup apparatus according to a first invention (claim 1) of the present application is an optical zoom means for optically zooming, an electronic zoom means for enlarging a digital image signal, and an image signal. It further comprises a face image recognition means for recognizing a human face portion and a zoom control means for controlling the optical zoom means and the electronic zoom means on the basis of the position and size information of the face area.

An image pickup apparatus according to a second invention (claim 2) of the present application, an optical zoom means for optically zooming, an electronic zoom means for enlarging a digital image signal, and a human face portion from the image signal. Face image recognition means for recognizing
A zoom control means for controlling the optical zoom means and the electronic zoom means based on the information of the position and size of the face area, and a display means for displaying an image before the image is enlarged by the electronic zoom means. is there.

An image pickup apparatus according to a third invention (claim 3) of the present application recognizes a registered subject from an image signal, an optical zoom means for optically zooming, an electronic zoom means for enlarging a digital image signal. Subject recognition means for
And a zoom control unit for controlling the optical zoom unit and the electronic zoom unit based on the recognized information on the position and size of the subject.

An image pickup apparatus according to a fourth aspect of the present invention (claim 4) is an optical zoom means for optically zooming, and a face image recognition means for recognizing a human face portion from an image signal.
An optical axis moving means for moving the optical axis and a zoom control means for controlling the optical zoom means and the optical axis moving means based on the information on the position and size of the face area are provided.

An image pickup apparatus according to a fifth aspect of the present invention (claim 5) is an optical zoom means for optically zooming, an electronic zoom means for enlarging a digital image signal, and a human face portion based on the image signal. Face image recognition means for recognizing
Zoom control means for controlling the optical zoom means and the electronic zoom means based on the information on the position and size of the face area, and a change for changing the detection area for detecting the focus position using the information on the position and size of the face area And means.

An image pickup device according to a sixth aspect of the present invention (claim 6) is a lens barrel for photographing an object, a photoelectric conversion means for converting an optical image from the lens barrel into an electric signal, and the lens barrel and the photoelectric conversion means. Holding unit for integrally holding the holding unit, a rotation supporting unit for rotatably supporting the holding unit in the panning direction and / or the pitching direction, an actuator unit for rotationally driving the rotation supporting unit, a holding unit and a device casing. It includes a detecting means for detecting a relative angle with the body, a returning means for returning the holding means to the reference position by a control value proportional to the detected relative angle, and a stationary detecting means for detecting a stationary state of the housing. When the detection means determines that the housing is stationary,
It is configured to increase the control gain to the returning means.

In the image pickup apparatus according to the seventh invention (claim 7) of the present application, the stationary state detecting means in the sixth invention determines that the stationary state is a stationary state because the angular velocity of the casing is small for a certain period of time, and the angular velocity of the casing is constant. When it has the above value, it is configured to judge that it is in a normal state.

An image pickup apparatus according to an eighth invention (claim 8) of the present application is a control gain according to the relative speed in the sixth invention,
It is configured to increase in proportion to the stationary time.

The image pickup apparatus according to the ninth invention (claim 9) of the present application is the same as the image pickup apparatus according to the sixth invention, in which when the stationary state is detected, it is displayed to the operator.

In the image pickup apparatus according to the tenth invention (claim 10) of the present application, the stillness detecting means in the sixth invention is determined to be in a stationary state when a fixing device such as a tripod is attached to the housing. It is composed.

An image pickup apparatus according to an eleventh invention (claim 11) of the present application is a lens barrel for photographing a subject, a photoelectric conversion unit for converting an optical image from the lens barrel into an electric signal, and the lens barrel and the photoelectric conversion unit. Holding unit for integrally holding the holding unit, a rotation supporting unit for rotatably supporting the holding unit in the panning direction and / or the pitching direction, an actuator unit for rotationally driving the rotation supporting unit, a holding unit and a device casing. Detecting means for detecting a relative angle with the body; first returning means for returning the holding means to a reference position by a control value proportional to the detected relative angle; and stillness detecting means for detecting a stationary state of the housing, Second holding means for mechanically fixing the holding means, and when the stationary state detecting means determines that the housing is stationary, the second returning means is operated to move the holding means to the reference position. Configured to return to It is a thing.

A motion vector detecting apparatus according to a twelfth invention (claim 12) of the present application is to provide a plurality of detection areas in a screen and obtain a correlation value at a predetermined deviation between the screens for each detection area. A means for obtaining a motion vector for each detection area from the correlation value, a means for moving a predetermined number of detection areas among a plurality of detection areas, and a motion vector for the entire screen using the motion vector for each area And means for doing so.

A motion vector detecting apparatus according to a thirteenth invention (claim 13) of the present application is to provide a plurality of detection areas in a screen and obtain a correlation value at a predetermined deviation between the screens for each detection area, A means for obtaining a motion vector for each detection area from the correlation value, a means for determining the reliability of the motion vector of each detection area, and a means for enabling movement of a predetermined number of detection areas among the plurality of detection areas, A means for deciding the motion vector of the entire screen using the motion vector of each area based on the judgment of reliability and moving the movable detection area is provided.

A motion vector detecting apparatus according to a fourteenth invention (claim 14) of the present application is to provide a plurality of detection areas in a screen and obtain a correlation value at a predetermined deviation between the screens for each detection area. A means for obtaining a motion vector for each detection area from the correlation value, a means for moving a predetermined number of detection areas among a plurality of detection areas, and a motion vector for the entire screen using the motion vector for each area And a means for doing so so that the correlation value detection points of the movable detection area do not overlap with the correlation value detection points of the other detection areas.

A motion vector detecting apparatus according to a fifteenth invention (claim 15) of the present application is to provide a plurality of detection areas in a screen and obtain a correlation value at a predetermined deviation between the screens for each detection area. A means for obtaining a motion vector for each detection area from the correlation value, a means for moving a predetermined number of detection areas among a plurality of detection areas, and a motion vector for the entire screen using the motion vector for each area In the tracking mode, a movable detection area is set as a tracking area, and a motion vector is detected only from the tracking area and the tracking area is moved according to the movement of the subject.

An image shake correcting apparatus according to a sixteenth invention (claim 16) of the present application is a motion vector detecting apparatus according to any one of the twelfth to fifteenth inventions, and a storage means for storing an image signal of one field or more, And a means for controlling the read position of the storage means based on the motion vector detected by the motion vector detecting device.

According to a seventeenth invention (claim 17) of the present application, an image shake correcting apparatus comprises a time axis fluctuation detecting means for detecting a time axis fluctuation of a video signal, a memory capable of writing and reading the video signal, and a video signal. An image shake detecting means for detecting image shake, and a write clock and a stable cycle for controlling the video signal write phase of the memory so as to correct the time axis fluctuation of the video signal according to the detection output of the time axis fluctuation detecting means. A clock generation unit that generates a read clock that controls the read phase of the memory, a memory write control unit that controls the write operation of the memory according to the timing of the write clock that the clock generation unit outputs, and a read that the clock generation unit outputs. The read phase of the memory is detected according to the clock timing, and it is detected by the image shake detection means. Memory read control means for reducing image shake by controlling read addresses of the memory in accordance with the shake amount.

In the image shake correction apparatus according to the eighteenth invention (claim 18) of the present application, the image shake detection means of the seventeenth invention uses 1 as a representative point among a plurality of specific pixels constituting the image.
In accordance with the representative point storage means for storing the field or for one frame, the correlation detecting means for detecting the correlation between the representative point and the surrounding pixels for one field or one frame, and the write clock output by the clock generating means. And a control means for controlling the writing phase of the representative point storage means and the phase of the correlation detection operation of the correlation detection means.

In the image shake correcting apparatus according to the nineteenth invention (claim 19) of the present application, the clock generating means of the seventeenth invention has a write clock generating means for generating a write clock and a frequency corresponding to the average frequency of the write clock. First read clock generating means for generating a clock having
Second read clock generating means for generating a clock having a constant frequency of crystal precision, and read clock for selecting and outputting either the output of the first read clock generating means or the output of the second read clock generating means. It has a selecting means and a read system synchronizing signal generating means for generating a synchronizing signal based on the output signal of the second read clock generating means.

An image shake correcting apparatus according to a twentieth invention of the present application (claim 20) is, in the seventeenth invention, including a video signal recording / reproducing means and recording / reproducing a video signal on the basis of a read clock outputted by the clock generating means. It is configured to control the reproducing operation of the means.

[0061]

In the image pickup apparatus of the first aspect of the invention, the zoom is adjusted so that the size of the face portion of the image becomes constant, so that a stable image can be obtained even if the subject moves.

In the image pickup apparatus of the second invention, the zoom is adjusted so that the size of the face portion in the image becomes constant, so that a stable image can be obtained even when the subject moves. By outputting the image before being magnified by the electronic image magnifying means in the electronic viewfinder, it becomes easy to follow a moving subject.

In the image pickup apparatus according to the third aspect of the present invention, since the subject is registered and the zoom is adjusted so that the size of the registered subject is constant, a stable image can be obtained for any subject. To be

In the image pickup apparatus of the fourth invention, since the zoom is adjusted so that the size of the face portion of the image becomes constant, a stable image can be obtained even if the subject moves and the optical Since the image is enlarged only by zooming, a high quality image can be obtained.

In the image pickup apparatus of the fifth invention, since the zoom is adjusted so that the size of the face portion of the image becomes constant, a stable image can be obtained even when the subject moves, and By using the information obtained by the image recognition means for changing the detection area of focus detection, an image in which the face is always in focus can be obtained.

In the image pickup apparatus of the sixth invention, when the still image pickup state is detected, the gain of the angle loop is increased in the control system of the lens barrel. As a result, the drive output applied to the actuator increases in proportion to the output value of the angle sensor.

In the image pickup apparatus according to the seventh aspect of the invention, the lens barrel is returned to the reference position when the angular velocity of the housing is below a certain level for a certain period of time.

In the image pickup apparatus according to the eighth aspect of the invention, when the stationary state is detected, the drive voltage for driving the actuator is
The component proportional to the relative angle between the lens barrel and the housing increases in proportion to the stationary time.

In the image pickup apparatus of the ninth invention, when the stationary state is detected, the change in the state of the image pickup apparatus is displayed through the viewfinder or the like.

In the image pickup apparatus of the tenth aspect of the invention, the lens barrel is returned when the fixing device such as a tripod is attached to the housing.

In the image pickup apparatus of the eleventh invention, when the stationary state is detected, the lens barrel is mechanically returned to the reference position.

In the motion vector detecting device of the twelfth invention, since the detection area is moved, for example, the user moves the detection area in consideration of the pattern, the position of the subject and the like.

In the motion vector detecting device of the thirteenth invention, since the detection region is moved according to the determination result of the reliability of the motion vector detected in each region, the detection region is automatically detected without the user's designation. Is set.

In the motion vector detecting device of the fourteenth invention, since the same detection point is not used when the movable area is moved and overlaps with the fixed area, the same detection point is duplicated. Instead of detecting, it detects from more detection points.

In the motion vector detecting device of the fifteenth invention, the tracking area in the tracking mode is also used as the movable area at the time of shake correction.

In the image shake correction apparatus of the sixteenth invention,
A blur correction device including a motion vector detection device having the actions of the twelfth invention to the fifteenth invention is realized.

In the image shake correction device of the seventeenth invention,
The phase of the write clock of the memory is adjusted according to the detection result of the time axis error of the input video signal, and the time axis of the video signal is corrected by reading this with a clock at fixed intervals, and at the same time the video signal is configured. The shake of the entire image is corrected by moving the read position of the memory according to the detection result of the shake of the entire image, and the shake of the screen due to the fluctuation of the time axis and the shake of the screen due to the shake of the hand are simultaneously removed.

In the image shake correction apparatus of the eighteenth invention,
A plurality of specific pixels among the pixels forming an image are stored as a representative point for one field or one frame, and the correlation between the representative point and the surrounding pixels is detected for one field or one frame. By controlling the write phase when storing the representative point and the phase of the correlation detection operation according to the output write clock, the positional relationship between the representative point and the pixel for which correlation detection is performed is unstable due to the time axis fluctuation of the video signal. To prevent becoming.

In the image shake correction apparatus of the nineteenth invention,
Either the first read clock, which is a clock having a frequency corresponding to the average frequency of the write clock, or the second read clock, which is a clock having a constant crystal precision frequency, is selected, and the second read clock is also selected. Generates a read-out synchronization signal based on the clock, and reproduces VT
When R cannot be externally synchronized, the read phase of the memory is controlled by the first read clock having the average frequency of the write clock frequencies, while when the external synchronization is possible, the highly accurate second read is performed. By controlling the read phase of the memory by the clock and controlling the reproducing operation of the reproducing VTR by the read system synchronizing signal, the conflict between the write operation and the read operation in the memory is prevented.

In the image shake correcting apparatus of the twentieth invention,
By controlling the reproduction operation of the video signal recording / reproducing means in combination with the video signal recording / reproducing means on the basis of the read clock output from the clock generating means, the phase of the reproducing operation in the video signal recording / reproducing means and the memory reading operation. Synchronize with the phase of.

[0081]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to the drawings showing the embodiments thereof.

(First Embodiment) FIG. 1 is a block diagram showing an image pickup apparatus according to the first embodiment of the present invention. In this example,
This is an example of applying auto zoom to a video movie.

In FIG. 1, 1, 21, 22, 5 to 12 are the same as those in FIG. 38 shown as a conventional example, and therefore their description is omitted. The light receiving surface 2 is a CCD element that converts an optical signal into an electric signal in this embodiment. The camera signal processing circuit 3 outputs a video signal V and a luminance signal Y from the electric signal obtained from the CCD element (light receiving surface) 2. The brightness signal Y is input to the focus detection circuit 4. The focus detection circuit 4 detects a high frequency component from the luminance signal Y. The high-frequency component of the luminance signal Y corresponds to the contrast of the screen, has the maximum value when the contrast is maximum, that is, when the focus is achieved, and decreases as the focus shifts. The focus lens control circuit 5 drives the focus lens 22 so that the focus evaluation value is maximized.

The video signal V output from the camera signal processing circuit 3 includes a luminance signal Y and color difference signals RY and BY, and is converted into a digital signal by the A / D converter 13. The face image recognition circuit 14 extracts a human face portion from the digital video signal obtained from the A / D converter 13,
Measure its size and position. Here, a specific operation of the face image recognition circuit 14 will be described with reference to FIG.

The input image is input to the skin color extraction circuit 24. In this skin color extraction circuit 24, R-Y is the x-axis and B-
Y is taken on the y-axis, the angle is θ1 to θ2, and the size is r1.
Those that satisfy the conditions from 1 to r2 are detected as the skin color. FIG. 3 shows the skin color range. Although the range of colors to be detected is fan-shaped, it may be a range of other shapes such as quadrangle and circle. Next, a binary image having a value of "1" for the part where the skin color is detected and "0" for the other parts is generated.

Next, the noise removing circuit 25 removes noise. An expansion / contraction process is performed to remove noise from the binarized data. For example, if the contraction process is performed 5 times and the expansion process is performed 5 times, a small region of 5 pixels or less is removed. Next, labeling is performed by the labeling circuit 26. This is a process of assigning a number to each extracted area.

Next, the characteristic amount calculation circuit 27 calculates the characteristic amount. First, the barycentric position Gi and the area A of each region
i, maximum width Wi, maximum height Hi, and perimeter Li are obtained.
Next, the aspect ratio Ri and the complexity Ci are obtained as the feature amounts by the following equations. Ri = Hi / Wi Ci = Li ² / Ai

Next, the judgment circuit 28 judges whether or not it is a face area. The judgment criteria are the following three points. (1) Ai> Amin A region having a certain area or more is extracted. (2) Rmin <Ri <Rmax A circumscribed quadrangle with an aspect ratio close to 1, that is, a quadrangle circumscribed with a square is extracted. (3) Ci <Cmax Extract a circle close to a circle.

An area satisfying the above three conditions is defined as a face image area. According to these three conditions, the face image can be recognized with a simple circuit and with high accuracy. In the determination circuit 28, the barycentric position G of the area extracted as the face image
Output i and area Ai.

The face image recognition circuit 14 is not limited to the above-mentioned method, and the face image area can be specified.
Any method can be used as long as its position and size can be obtained.

The zoom control circuit 15 controls the zoom lens control circuit 9 and the electronic zoom circuit 16 based on the size and position information of the face portion in the image obtained from the face image recognition circuit 14.

The operation of the zoom control circuit 15 will be described with reference to FIG. First, it is assumed that the image shown in 31 is obtained. In the face image recognition circuit 14, 32 out of 31 images
The portion indicated by is recognized as a face, and its size and position are obtained. Zoom control circuit that obtains information about face size and position
At 15, the zoom lens control circuit 9 is instructed to move the zoom lens 21 to the tele side so that the face area does not leave the frame and the face area is as close to the size of the bust shot as possible. Here, the information on the size and position of the face is always obtained from the face image recognition circuit 14, and the control is performed so as to be close to the information on the size and position of the face at the time of a bust shot given in advance. There is. In this way, the zoom lens 21 is moved, and an image 34 in which 33 of the 31 images are zoomed up is obtained. In 34, in order to obtain a bust shot image based on the information of the face area 32, the area of 35 is enlarged by the electronic zoom circuit 16 to obtain the bust shot image.
Get 36. In summary, based on the information of the face area 32 in 31, the optical zoom is used to enlarge 33 to 34, and the electronic zoom is used to enlarge 35 to 36 to obtain a bust shot image.

The zoom control circuit 15 also controls the enlargement ratios of the optical zoom and the electronic zoom. First, the movement of the face area is detected. This is a face image recognition circuit
Regarding the position information of the face area obtained from 14, the difference from the value on the previous screen is calculated. If the value is large, it is determined that the movement is strong, and if the value is small, it is determined that the movement is small. When the subject moves rapidly, the enlargement ratio of the optical zoom is reduced and the enlargement ratio of the electronic zoom is increased. As a result, the range in which the clipping frame of the electronic zoom image can move becomes wider, that is, the correction range becomes wider, and a moving subject can be tracked. Further, when the subject is near the center of the screen and the movement is small, the enlargement ratio of the enlargement by the optical zoom is increased and the enlargement ratio of the electronic zoom is reduced. As a result, the image quality deterioration of the electronic zoom can be reduced and the image quality can be improved.

When this method is used, the size of the face is used as a condition for determining the zoom magnification, so the size of the face in the screen is always constant, and the size becomes larger or smaller as in the conventional example. There is nothing to do. Further, even when the subject is not in the center of the screen, the clipping position of the electronic zoom image as shown at 35 in FIG. 4 is set to the position with the face, so that a bust shot as shown at 36 can be obtained. . When the subject moves, the clipping position of the electronic zoom image moves according to the movement of the subject, and a stable bust shot can always be obtained. Further, from these facts, there is also an effect of correcting camera shake.

In this embodiment, a bust shot is assumed as the final composition, but the composition is not limited to this, and may be a composition of a person, such as a shot of the whole body or only the face is placed in the lower left of the screen. Anything will do. It is also possible to preset these plural compositions and allow the user to select and use them.

As described above, in the first embodiment, a stable image can be obtained even when the subject moves, and the subject can be photographed with a fixed composition only by pointing the camera in the approximate direction. .

(Second Embodiment) Next, the screen of the electronic viewfinder (EVF) of the image pickup apparatus shown in the first embodiment is made switchable. FIG. 5 is a block diagram showing an image pickup apparatus according to the second embodiment of the present invention. Comparing FIG. 5 with FIG. 1, it is the same except that an EVF image control circuit 17 and an EVF 18 are added.

The operation of the second embodiment will be described. EV
The F image control circuit 17 outputs the image data before electronic zooming from the output of the A / D converter 13, the range in which electronic zooming is performed from the zoom control circuit 15, and the electronic zooming of the image data after electronic zooming. Obtained from circuit 16. To explain this with reference to FIG. 4, the image before electronic zoom is
34, the range of electronic zoom is 35, and the image after electronic zoom is 36. Here, in the first embodiment, the image 36 after electronic zoom is output to the EVF 18. In the second embodiment,
Image showing the range of electronic zoom in the image before electronic zoom,
That is, an image in which 35 frames are displayed in 34 images can be output. Then, two images, an image before the electronic zoom and an image after the electronic zoom, can be selected as the output of the EVF 18. By this,
In the case of a subject that moves rapidly, the image before the electronic zoom is output from the EVF 18, which makes it easier to follow the moving subject.

As described above, in the second embodiment, similar to the first embodiment, a stable image can be obtained even when the subject moves, and the movement of the subject can be confirmed.

(Third Embodiment) Next, an embodiment will be described in which, in the image pickup apparatus shown in the first embodiment, recognition of a face image is performed. FIG. 6 is a block diagram showing an image pickup apparatus according to the third embodiment of the present invention. Comparing FIG. 6 with FIG. 1, the face image recognition circuit 14 is replaced with a subject recognition circuit 19, and a subject registration circuit 20
It is the same except that is added.

The operation of the third embodiment will be described. First
In the embodiment, the recognition symmetry is limited to the human face,
The third embodiment does not limit the symmetry of recognition. First,
Aim the camera at the subject you want auto zoom to work on,
Take the desired composition. Here, the features (shape, color, etc.) and composition of the subject are registered in the subject recognition circuit 19.

The subject registration operation will be described with reference to FIGS. 7 and 8. FIG. 7 is a diagram showing a state of subject recognition in the third embodiment of the present invention. First, the composition of the subject is determined. Then, the registration point 41 is moved to the location of the subject 42 to which auto zoom is to be applied. In this state, press the registration switch. At this time, the image is stored in the memory, and at the same time, the subject registration circuit 20 calculates the average value of the color difference signals in the registration point 41. For example, it is assumed that the average value of the obtained color difference signals is 43 points shown in FIG. Here, 44 is set as the detection color area. In the color difference signal at the registration point, the angle is ± 30 ° and the size is ± 20 with respect to the average value 43 of the color difference signal. Next, with respect to the image of the subject registered in the memory, the color area included in the detected color area 44 is extracted from the image recorded in the memory. From the extracted areas, the area including the registration point is recognized as a subject, and its shape and position are recorded in the memory. Furthermore, the area, aspect ratio, and complexity, which are the feature quantities of the recognized subject, are calculated and stored.

When a subject is registered, what is performed in the first embodiment is not performed on the human face but is performed on the registered subject. This makes it possible to apply auto-zoom to any subject with any composition.

Regarding the subject registration, the feature amount may be other parameters such as texture and motion information, or a combination thereof. Also for color detection, the subject may be registered as a combination of a plurality of color components having peaks instead of the average value.

In the third embodiment, similar to the first embodiment, a stable image can be obtained even when the subject moves, and a stable image can be obtained for an arbitrary subject.

(Fourth Embodiment) Next, an embodiment will be described in which the image pickup apparatus shown in the first embodiment is used to track an object by moving the lens barrel. FIG. 9 is a block diagram showing an image pickup apparatus according to the fourth embodiment of the present invention. Comparing FIG. 9 with FIG. 1, there is no electronic zoom circuit 16 and a lens barrel control circuit 30 is newly added. Further, the lens barrel 1 is controlled by the lens barrel control circuit 30 and its direction can be changed.

The operation of the fourth embodiment will be described. First
In the embodiment, the electronic zoom is performed by the electronic zoom circuit 16 after the optical zoom by the zoom lens 21.
Explaining with reference to FIG. 4, enlargement from the image 31 to the image 34 is performed by the optical zoom, and enlargement from the image 34 to the image 36 is performed by the electronic zoom. In the fourth embodiment, electronic zoom is not performed. The face image recognition circuit 14 measures the position and size of the face image from the digital image data obtained by the A / D converter 13, as in the first embodiment. It is assumed that the face image recognition circuit 14 obtains an image as shown at 31 in FIG. At this time, the position of the human face area 32 is output to the lens barrel control circuit 30. The lens barrel control circuit 30 controls the direction of the lens barrel 1 so that the human face area 32 is located at the center of the screen. Further, according to the information of the size of the human face of the zoom control circuit 15, the zoom lens control circuit 9 moves the zoom lens 21 so that the image of the subject becomes a bust shot. By the above operation, 36 images are obtained from the image 31 of FIG. By using this method, the same effect as that of the first embodiment can be obtained. Further, since the electronic zoom is not used in the fourth embodiment, the image quality does not deteriorate.
Further, in the fourth embodiment, the image of the subject is moved to the center of the screen by moving the lens barrel, but the same effect can be obtained by using a lens that moves the optical axis.

In the image pickup apparatus of the fourth embodiment, similar to the image pickup apparatus of the first embodiment, a stable image can be obtained even when the subject moves, and an image without image quality deterioration can be obtained.

(Fifth Embodiment) Next, an embodiment in which autofocus is taken into consideration in the image pickup apparatus shown in the first embodiment will be described. FIG. 10 is a block diagram showing an image pickup apparatus according to the fifth embodiment of the present invention. Comparing FIG. 10 with FIG. 1, the constituent elements are the same. However, the information on the position and size of the human face obtained by the face image recognition circuit 14 is output to the focus detection circuit 4.

Focus detection for conventional focusing has been performed using data of the entire screen or a fixed part of the screen. As a result, the background may be in focus rather than the subject that is actually being focused on. If this happens during auto-zoom, it may happen that the zoomed-in subject is out of focus. The focus detection circuit 4 of the fifth embodiment detects the focus of the face portion based on the face position and size information from the face image recognition circuit 14. Explaining with reference to the image 31 of FIG. 4, focus detection is performed on the 32 parts detected as a face, and the focus lens 22 is moved. This makes it possible to obtain an image in which the face portion is always focused.

In the image pickup apparatus of the fifth embodiment, similar to the image pickup apparatus of the first embodiment, it is possible to obtain a stable image even when the subject moves, and it is possible to obtain an image always focused.

(Sixth Embodiment) FIG. 11 is a block diagram showing the image pickup apparatus according to the sixth embodiment of the present invention. In the figure, 51
Is a lens barrel (optical system), 52 is a gimbal mechanism that supports the lens barrel 51, 53 is an actuator that drives the lens barrel 51, and 54 is the lens barrel 51.
And an angle sensor that detects a relative angle between the housing and a housing (not shown),
55 is an angular velocity sensor for detecting the angular velocity generated in the lens barrel 51,
56 is a variable amplification circuit that multiplies the gain of the angle sensor 54 by which it can be changed, 57 is a variable amplification circuit that multiplies the output of the angular velocity sensor 55 by which it can be changed, and 58 is the angle sensor 54 and the angular velocity sensor 55. A stationary state detection circuit that determines the stationary state of the lens barrel 51 from the output, and 59 is a variable amplification circuit 5
6, an adder for adding the outputs of 57, 60: an actuator drive circuit for driving the actuator 53 from the output of the adder 59, 61: an image pickup device for picking up an image of a subject through the lens barrel 51, 62
Is a signal processing circuit that processes a video signal output from the image sensor 61, and 63 is a display device that displays an image state.

In FIG. 11, only the actuator, the sensor and the control circuit for rotating the lens barrel 51 in the horizontal direction (yaw direction) are shown, but the same structure is applied in the vertical direction (pitch direction). Has been realized.

A large number of lens groups (not shown) are attached to the lens barrel 51,
It is composed of an image pickup device 61 such as a CCD or an image pickup tube. The lens barrel 51 forms an image of the subject on the image pickup device 61, and the image pickup device 61 converts the image into an electric signal. Lens barrel
Reference numeral 51 denotes an image pickup device for forming an image of a subject on the image pickup device 61.
The video signal is output from 61. This video signal is processed by the signal processing circuit 62, converted into an NTSC video signal and output.

On the other hand, the lens barrel 51 is rotatably supported by the gimbal mechanism 52. A specific configuration example of this mechanism is shown in FIG. In the figure, 71 is a first support and 72 is a second support.
Is a support body, 73 is a first rotation axis, 74 is a second rotation axis orthogonal to the first rotation axis 73, 75 is a first actuator, and 76 is a second actuator.

The lens barrel 51 is rotatably supported inside the second supporting member 72 about the second rotating shaft 74 as a central axis, and is driven in the pitch direction by driving the second actuator 76. Rotation can be controlled. Also, the second support
72 is rotatably supported inside the first support body 71 about the first rotation shaft 73 as a center axis. By driving the first actuator 75, the lens barrel 51 is rotated in the yaw direction. The movement can be controlled. The first support 71 is fixed inside the housing (not shown) of the imaging device.

With such a structure, the lens barrel 51
By driving the first and second actuators 75 and 76 respectively with the outputs of the control circuit, they can be rotated in the yaw direction and the pitch direction with respect to the first support 71.

In FIG. 12, the yaw direction is the first rotation axis,
Although the pitch direction is illustrated as the second rotation shaft, the pitch direction is not limited thereto, and the pitch direction can be realized as the first rotation shaft and the yaw direction as the second rotation shaft.

Next, the camera shake correction operation of the image pickup apparatus configured as described above will be described in the yaw direction. In addition,
The following description can be similarly applied to a device configuration in the pitch direction not shown in FIG.

Since the photographer holds the casing of the apparatus or the grip portion attached to the casing for photographing, vibration not intended by the photographer is generated in the casing due to camera shake of the photographer. This vibration is transmitted to the lens barrel 51, the angular velocity is detected by the angular velocity sensor 55, and converted into a corresponding electric signal ω. This signal ω is amplified (angular velocity loop gain Kω) by the variable amplifier circuit 57 and input to the adder 59.

On the other hand, the relative angle between the lens barrel 51 and the housing is detected by the angle sensor 54 and converted into a corresponding electric signal θ. This signal θ is amplified by the variable amplification circuit 56 (angle loop gain Kθ) and input to the adder 59.

The adder 59 adds the outputs of the variable amplifier circuits 56 and 57 to calculate a control output value, and outputs the control output value.
Enter in 60. The actuator drive circuit 60 drives the actuator 53 with this control output value.

FIG. 13 is a block diagram of the control system of FIG.
In the figure, 81 is an angular loop gain Kθ, 82 is an actuator drive gain, 83 is a mechanism system representing the lens barrel 51 and the gimbal mechanism 52, 84 is an angular velocity loop gain Kω, and 85 is a differential element.

In the normal mode (normal camera shake correction state), the variable amplifier circuit 56 gives a constant Kθ given in advance as a multiplication coefficient Kθ to be multiplied by the output value θ of the angle sensor 54.
Has 0.

The stationary state detection circuit 58 calculates the angular velocity ω 0 of the housing in the inertial coordinate system from the angular velocity ω and the relative angle θ. ω 0 = dθ / dt-ω

Then, if the angular velocity of this housing is equal to or less than a certain value for a certain time given in advance, it is judged that the image pickup apparatus is in a stationary state. That is, the stationary state detection circuit 58 temporally differentiates the output value θ of the angle sensor 54,
Subtract the output value ω of the angular velocity sensor 55 from the calculation result,
Obtain the angular velocity ω 0 of the housing.

The stationary state detection circuit 58 constantly monitors this value ω 0 and measures the time which is equal to or less than the preset threshold value ω s. That is, when ω0 <ωs continues to be satisfied for a certain time or more, it is determined that the stationary state is established, and the normal mode coefficient Kθ0 of the angle loop is set to the stationary mode coefficient Kθs (K
θs> Kθ0).

As described above, in this embodiment, the control system has two states (modes) of the normal mode and the stationary mode.
And changes the gain of the control system according to the state of the image pickup apparatus. The operation of the stationary state detection circuit 58 that detects the state in this embodiment will be described below with reference to the flowchart of FIG.

In the initial state, the normal mode is set (step S1), and the camera shake correction function is prioritized in operation. At this time, the counter count is 0 (step S2). The case angular velocity ω 0 is calculated (step S
3) After that, the calculated ω0 is compared with a preset threshold value (a small value regarded as a stationary state) ωs (step S4). If ω0 is large here, the camera shake correction should be continued, and the process returns to step S1.

However, if ω0 is smaller than ωs, the image pickup device may be in a stationary state. Therefore,
The counter count is incremented by +1 (step S5),
Measure the duration. Then, the value of the counter is compared with a threshold value TIME given in advance (step S
6). If the value of the counter exceeds the threshold value TIME, it is determined that the image pickup apparatus has been in the stationary state for a sufficiently long time, so the mode shifts to the stationary mode (step S
7). Although not shown in the flowchart of FIG. 14, the value of the angle loop gain Kθ is changed from Kθ0 to Kθs at this time. As a result, the component included in the drive output of the actuator, which is proportional to the relative angle between the central axis of the lens barrel and the central axis of the housing, increases, and the origin return performance improves. Note that the figure
In the flowchart of FIG. 14, the variable value overflow processing and the like are omitted for simplification.

Further, when the gain of the angle loop is rapidly increased in the stationary mode, the operation of the lens barrel suddenly changes against the operator's intention, which causes a problem in operability.
Therefore, a smooth operation can be obtained by interpolating the finally set gain Kθs and the initial gain value Kθ0 and gradually increasing the gain while the stationary mode continues. In FIG. 15, the horizontal axis represents the elapsed time in the stationary mode, and the vertical axis represents the gain of the angle loop. As shown in this figure, a sudden change in the control system can be prevented by changing the gain with respect to the elapsed time in the stationary mode.

Further, as the above-mentioned interpolation method, the operability can be further improved by giving a function having an arbitrary shape instead of the simple linear interpolation. In FIG. 16, the horizontal axis represents the elapsed time in the stationary mode, and the vertical axis represents the gain of the angle loop. In this figure, the gain is set as a quadratic function of time.

Further, as shown in FIG. 11, by displaying the state on a display device 63 such as a viewfinder, it is possible to reduce the discomfort felt by the operator.

Even in the stationary mode, the monitoring of the cabinet angular velocity is continued in steps S3 and S4. If the housing angular velocity ω 0 exceeds the threshold ω s, the normal mode is immediately entered.

(Seventh Embodiment) FIG. 17 is a block diagram showing the image pickup apparatus according to the seventh embodiment of the present invention. In this embodiment,
The operation is the same as that of the sixth embodiment except the operation of the stationary state detection circuit 58, and FIG. 18 is a flow chart for explaining the algorithm at the time of mode switching in the seventh embodiment.

In FIG. 18, the operation mode is initialized (step S1) and the timer is initialized (step S2).
Then, the stationary state detection circuit 58 compares the output θ of the angle sensor 54, which detects the relative angle between the lens barrel 51 and the housing, with a small threshold value θs given in advance (step S
8). If the output of the angle sensor 54 is smaller than the threshold value, the duration is measured (step S5). Then, when a certain time or more has passed (step S6),
The image pickup apparatus judges that it is in the stationary state, switches the operation mode as in the sixth embodiment, and changes the gain of the control system (step S7).

(Eighth Embodiment) FIG. 19 is a block diagram showing the image pickup apparatus according to the eighth embodiment of the present invention. This embodiment is the same as the sixth embodiment except for the operation of the stationary state detection circuit 58, and FIG. 20 is a flow chart for explaining the algorithm at the time of mode switching in the eighth embodiment.

In FIG. 20, after initialization (steps S1 and S2), the stationary state detection circuit 58 sets the control value v of the actuator 53 for driving the lens barrel 51 to a threshold value vs given in advance. Compare (step S9). When the control value v is smaller than the threshold value vs, the duration is measured as in the sixth embodiment (step S5), and when a certain time has passed (step S6), it is determined that the control value v is stationary.
The subsequent processing is the same as in the sixth embodiment.

(Ninth Embodiment) FIG. 21 is a block diagram showing the image pickup apparatus according to the ninth embodiment of the present invention. In this embodiment,
In addition to the configuration of the sixth embodiment, a sensor 91 such as a micro switch is provided at the tripod mounting position of the casing 93 of the image pickup device,
It is configured to detect the mounting state of 92. The stationary state detection circuit 58 monitors the output of the sensor 91, and when the tripod 92 is attached, determines that the stationary imaging state. The subsequent processing is the same as in the sixth embodiment.

(Tenth Embodiment) FIG. 22 is a block diagram showing the image pickup apparatus according to the tenth embodiment of the present invention. In this embodiment,
In addition to the configuration of the sixth embodiment, a contact detection sensor 94 such as a micro switch is provided below the housing 93. When the sensor 94 detects a contact, it is considered that the imaging device is placed on a table or the like. Therefore, the stationary state detection circuit
The output of the sensor 94 allows the 58 to determine that the housing is stationary. However, a plurality of sensors 94 (two in this example) are provided so that the operator does not accidentally touch them to operate, and the logical product of the outputs of these sensors 94 is obtained by the AND circuit 95. Judge by. The subsequent processing is the same as in the sixth embodiment.

(Eleventh Embodiment) In the sixth to tenth embodiments described above, a method of increasing the gain of the angle loop is used to return the lens barrel 51 to the reference position. The same effect can be obtained by various fixing devices.

FIG. 23 is a block diagram showing the image pickup apparatus in the 11th embodiment of the present invention. When the stationary state detection circuit 58 detects the stationary state of the image pickup device, the fixing device 96 of the lens barrel 51 can fix the lens barrel 51 to the reference position. Further, by stopping the power supply to the drive mechanism of the lens barrel 51 during the fixing, the power consumption can be suppressed.

The mechanism for returning the lens barrel to the reference position in the sixth to eleventh embodiments can be used for fixing the lens barrel when the power is turned off or when the camera shake correction function is not used.

As described above, according to the image pickup apparatus of the sixth to eleventh embodiments, it is possible to obtain a sufficient shake correction ability even in a control system having a relatively low resolution. Further, since the resolution of the control system can be set low with respect to the required camera shake correction ability, the hardware scale can be reduced. Also, because the control system used for camera shake correction is shared, it is possible to prevent an increase in hardware scale,
In addition, high-speed mode switching is possible.

(Twelfth Embodiment) FIG. 24 is a block circuit diagram of a motion vector detecting apparatus according to the twelfth embodiment of the present invention. An example will be described which is composed of five fixed detection areas (HL, HR, LL, LR) and one movable area (CT). In FIG. 24, the same reference numerals as those in FIG. 40 indicate the same parts, and 119 is a CT area offset setting circuit that gives an offset from the reference position to the CT area, which is a movable area, from the reliability determination result of each area. is there.

FIG. 25 is a block circuit diagram showing an embodiment of the CT area offset setting circuit 119. Actually, there are two block counters in the horizontal direction and the vertical direction.
For convenience, only one block counter 121 is used. In the case of another fixed area, for example, the HL area, the area gate pulse can be created by decoding the block counter with the HL area start decoding circuit 122 and the HL area end decoding circuit 123, and setting and resetting them respectively. In the case of the CT area, the block counter value and the offset value from the offset value generating circuit 127 are added by the adding circuit 124, and the addition result is similarly obtained in the CT area start decoding circuit 125 and the CT area end decoding circuit.
A movable area gate pulse can be created by decoding with 126.

Next, the operation will be described. For example, Figure 26
If a landscape such as that shown in Figure 4 is taken from a car, the values of the representative points of the previous screen are stored in the representative point memory 101, and the absolute value calculation circuit 102 determines the absolute values of the pixels around the representative points of the current screen. It is calculated. The calculated absolute value passes through each gate circuit 103, 106, 109, 112, 115 and is integrated by each cumulative addition circuit 104, 107, 110, 113, 116 to obtain the motion detection area of each motion detection area. Only the timing is cumulatively added. The reliability of each cumulative addition value is judged by each reliability judgment circuit 105, 108, 111, 114, 117. Basically, when the reliability is high, the minimum position of the cumulative addition value is set as the motion vector. .

In FIG. 26, in the HL area 131, most of the mountains are occupied and there are some buildings. in this way,
When there is no change in the brightness level on the mountain, the reliability is low as shown in Fig. 42 (b), but the reliability is low due to the presence of buildings and the difference in the way of trees, and the reliability is low. It is determined that there is. In the HR area 132, one side is occupied by the sky. In this case, as shown in FIG. 42 (b), the reliability is judged to be low. In the LL area 133, since natural images of buildings, trees, etc. are shown, it becomes as shown in FIG. 42 (a), and it is judged that the reliability is high. The LR area 134 and the CT area 135 are areas where the sea occupies most of the area. When there is no wave, the area is close to that shown in FIG.
Although it is close to (c) and (d), the reliability is judged to be low in any case.

Therefore, it is determined that the HL area and the LL area have high reliability and the HR area, the LR area, and the CT area have low reliability. Therefore, the final motion vector is determined from the HL area and the LL area, and if the number of representative points in one area is 30, the motion vector will be detected from 60 representative points.

At this time, from the reliability determination result, the left side of the screen has higher reliability, so it is considered that moving the CT area to the left side is likely to improve the reliability. To be Therefore, for example, according to the algorithm shown in FIG. 28, since the reliability of the CT area and the LL area is high, (x, y) = (− 2.0, 0), and in the next screen, move the CT area two steps to the left. To move. On the next screen, by repeating the same algorithm, if the reliability determination result does not change, the screen is moved to the left by two steps in units of screen, and the movement is stopped when the reliability becomes high.

In this state, the reliability of the CT area is also judged to be high, so the CT area is added to the HL area and the LL area to determine the final motion vector. At this time, the representative score used for the final decision is 90 points, and the detection accuracy is improved compared to the initial 60 points.

As described above, since the detection area can be moved, for example, when the user specifies the detection area in consideration of the pattern, the position of the subject, etc., the operation is performed so as to eliminate the blurring of that portion. be able to. Further, since the detection area is moved according to the determination result of the reliability of the motion vector detected in each area, the detection area can be automatically set to an ideal pattern or subject without the user's designation.

(Thirteenth Embodiment) Next, the thirteenth embodiment is arranged so that the detection positions (representative points) do not overlap when the movable area is moved.
Examples will be described. For example, as shown in FIG. 25, the gate circuit 128 calculates the logical product of the gate pulse of the fixed area and the least significant bit of the block counter, and the gate circuit 129
And the logical product of the gate pulse of the movable area and the inversion of the least significant bit of the block town. By inserting this process both horizontally and vertically, the representative points can be arranged as shown in FIGS. 27 (a) and 27 (b). 27 (a), (b)
In the figure, a right-hatched area 135 indicates a fixed area detection block position, and a left-hatched area 136 indicates a moving area detection block position. In this case, when the areas are in contact with each other as shown in FIG. 27A or when the areas are overlapped as shown in FIG. In comparison, the final motion vector can be detected from many positions.

Therefore, since the same detection point is not used when the movable area is moved and overlaps with the fixed area, the detection is not performed redundantly at the same detection point but from more detection points. It can be detected and the detection accuracy is improved.

Next, the operation in the tracking mode in the 12th or 13th embodiment will be described. Here, the tracking mode means that the distance measuring area for auto focus or the light measuring area for auto iris is moved following the movement of the subject. In the motion vector detection in the image shake correction mode, as described above, the position of the movable area is determined and moved based on the reliability determination result of all areas. In the tracking mode, the position of the movable area is determined and moved by the motion vector itself of the movable area. Therefore, the tracking area in the tracking mode and the movable area at the time of image shake correction can be configured to be shared, and both functions can be realized without increasing the circuit scale.

Although the twelfth or thirteenth embodiment described above is an example of the representative point matching method, an all-point matching method for obtaining the correlation for all pixels or another detection method may be used. Moreover, although the case where the fixed area is four and the movable area is one has been described, the number of these areas may be any number. Also, as an example of reliability judgment algorithm, an example in which the change of the pattern is too small, the change is periodic, and the change is too large has been explained, but these temporal and positional changes are used. Various algorithms are conceivable, such as determining the intrusion of a moving object. Further, although the movable area has been described as an example in which it is moved by several steps, it may be moved by a large amount. Further, in the algorithm of FIG. 28, when (x, y) = (0, 0), CT
I explained an example of not moving the area, but at this time, CT
The area may be moved in an oscillating manner.

(Fourteenth Embodiment) An image shake correcting apparatus using the motion vector detecting apparatus configured as described above will be described. FIG. 29 is a block diagram showing the structure of the image shake correction apparatus in the fourteenth embodiment. In FIG. 29, reference numeral 137 is a video memory for storing at least one field video signal, 138 is a motion vector detecting device of the twelfth or thirteenth embodiment described above, and 139 is a control circuit for controlling the read position in the video memory 137. .

When a video signal is input to the video memory 137, 1
Video signals of fields and above are stored. The video signal is also input to the motion vector detection device 138, the motion vector is detected as described above, and the detected motion vector is output to the control circuit 139. The control circuit 139 is
Based on this motion vector, the read position of the video signal of the video memory 137 is controlled. Therefore, it is possible to realize an image shake correction device that corrects the shake for any object without feeling uncomfortable.

(Fifteenth Embodiment) FIG. 30 is a block diagram showing the overall structure of an image shake correction apparatus according to the fifteenth embodiment of the present invention. In FIG. 30, 151 and 152 are Y / C separate video signal input terminals to which a chroma signal and a luminance signal are respectively input, 153 is a composite video signal input terminal, and 154 is a composite video signal input to the input terminal 153. A Y / C separation circuit for separating a chroma signal and a luminance signal, 155 is an input selector switch for selecting either the video signal from the input terminals 151 and 152 or the video signal from the Y / C separation circuit 154, 156 Is a sync separation circuit for separating the sync signal from the luminance signal from the input selector switch 155.

Further, 157 is a clock generation circuit for generating a write system clock WCLK and a read system clock RCLK according to the synchronization signal output from the synchronization separation circuit 156,
Reference numeral 158 is a read system synchronization signal RSYN based on the read system clock RCLK output from the clock generation circuit 157.
C is a synchronizing signal generating circuit, 159 is a divide-by-4 circuit that divides the write-system clock WCLK output by the clock generating circuit 157 into four, and 160 is an input selector switch 155 based on the output signal of the divide-by-4 circuit 159. It is a color decoder that converts the output chroma signal into two color difference signals RY and BY.

Further, 161, 162, 163 are color decoders 16
A / D converter for converting the RY and BY signals output by 0 and the luminance signal output by the input selector switch 155 into digital signals at the timing of the write system clock WCLK output by the clock generation circuit 157. , 164, 1
65 and 166 are R output from the A / D converters 161, 162 and 163
The -Y signal, the BY signal, and the luminance signal are written at the timing of the write system clock WCLK output from the clock generation circuit 157, and after one field period, read at the timing of the read system clock RCLK. .

Reference numeral 167 is an image shake detecting section for detecting the motion vector of the entire image for each field from the digitally converted luminance signal output from the A / D converter 163, and 168
Is a write system memory controller for controlling the write operation of the field memories 164, 165, 166 by the write system clock WCLK output by the clock generation circuit 157 and the synchronization signal output by the sync separation circuit 156, and 169 is generated by the clock generation circuit 157. Depending on the read system clock RCLK and the output of the image shake detection unit 167, the field memory
It is a read-out type memory controller that controls the reading of 164, 165 and 166.

170, 171, and 172 are field memories.
Interpolation enlargement circuit that enlarges the output signals of 164, 165, and 166 to the original screen size, and 173, 174, and 175 are interpolation enlargement circuits.
A D / A converter that converts the outputs of 170, 171, and 172 into analog signals, and 176 is a sync signal replacement circuit that adds the sync signal output by the sync signal generation circuit 158 to the luminance signal output by the D / A converter 175. is there.

Further, 177 is a parallel modulation circuit for converting the two color difference signals output from the D / A converters 173, 174 into a chroma signal of NTSC standard, and 178, 179 are Y / Y for outputting a chroma signal and a luminance signal, respectively. C separate video signal output terminal, 180 is a Y / C mixing circuit for mixing the chroma signal output from the parallel modulation circuit 177 and the luminance signal output from the sync signal replacement circuit 176 to generate a composite video signal, 18
An output terminal 1 outputs the output signal of the Y / C mixing circuit 180 to the outside.

Next, the operation of this embodiment will be described.
If the video signal to be shake-corrected is a Y / C separate signal, it is input to the input terminals 151 and 152, and if it is a composite video signal, it is input to the input terminal 153. Further, the video signal input to the input terminal 153 is separated into a luminance signal and a chrominance signal by the Y / C separation circuit 154. Input selector switch 155
Is the luminance signal and color signal from the input terminals 151, 152 and Y / C
Either the luminance signal or the chrominance signal output from the separation circuit 154 is selected. The sync separation circuit 156 has an input selector switch 15
Extract the sync signal from the luminance signal selected by 5,
It is output as a write system horizontal synchronizing signal WHD.

The clock generation circuit 157 is a synchronization separation circuit 15
The write system clock WCLK and the read system clock RCL according to the write system horizontal synchronizing signal WHD output by
Generate K and. The write system clock WCLK is a clock modulated by the time axis error detected from the write system horizontal synchronization signal WHD, and is the A / D converter 161, 162, 163.
, The reference clock of the image shake detection unit 167 and the write system memory controller 168, and the read system clock R
CLK is a synchronizing signal generation circuit 158, a read system memory controller 169, interpolation expansion circuits 170, 171, 172, and D.
It serves as a reference clock for the A / A converters 173, 174, and 175. The average frequency of the write system clock WCLK and the read system clock RCLK is four times the color subcarrier frequency fsc.

FIG. 31 shows a concrete configuration of the clock generation circuit 157. In FIG. 31, 182 is a first phase-locked loop (hereinafter referred to as “PLL ")).

FIG. 32 is a block circuit diagram showing a specific structure of the first PLL 182. In FIG. 32, 186 is a phase comparator that outputs a voltage according to the phase difference between the two input signals, and 187 is a voltage controlled oscillator that generates a signal whose frequency changes according to the voltage output by the phase comparator 186. Reference numeral 188 denotes a 910 frequency divider circuit for dividing the output signal of the voltage controlled oscillator 187 by 910. The phase comparator 186 uses the 910 divider circuit 188
The output signal of the voltage controlled oscillator 187 divided by 10 is compared with the phase of the write system horizontal synchronizing signal WHD output by the sync separation circuit 156, and a voltage corresponding to the phase difference is output to the voltage controlled oscillator 187. And constitutes a PLL driven by the write-system horizontal synchronization signal WHD.

The frequency dividing ratio of the 910 frequency dividing circuit 188 is determined to be 910 because the relationship between the frequency fsc of the color subcarrier and the horizontal synchronizing frequency fH is defined as fsc = 455/2 × fH in the NTSC video signal. At this time, an output signal having a frequency 910 times the frequency fHW of the write system horizontal synchronizing signal WHD, that is, a frequency four times the frequency fscW of the color subcarrier of the input video signal can be obtained from the voltage controlled oscillator 187.

Further, in FIG. 31, reference numeral 183 denotes a second PLL which outputs a signal obtained by multiplying the average frequency / fHW of the write system horizontal synchronizing signal WHD by 910, and a concrete configuration example thereof is shown in FIG. In FIG. 33, 189 is a phase comparator, 190 is a low-pass filter (hereinafter referred to as “LPF”) that extracts and outputs the low-frequency component of the output signal of the phase comparator 189, 191
Is a voltage controlled oscillator that generates a signal having a frequency corresponding to the output voltage of the LPF 190, and 192 is a 910 frequency divider circuit that divides the output signal of the voltage controlled oscillator 191 by 910. Phase comparator 189
Generates a voltage signal corresponding to the phase difference between the write-system horizontal synchronizing signal WHD and the output signal of the 910 frequency divider circuit 192.
Output to F190. Therefore, the phase comparator 189 and the LPF 19
0, the voltage controlled oscillator 191, and the 910 frequency divider 192 constitute a PLL driven by the write system horizontal synchronizing signal WHD, and the time constant of the loop is determined by the characteristics of the LPF 190.

Since the phase comparator 189 compares the phase of the output of the voltage controlled oscillator 191 with the write system horizontal synchronizing signal WHD at the division ratio of 910, the output signal of the voltage controlled oscillator 191 is the write system horizontal synchronizing signal WHD. The frequency is 910 times the frequency fHW. Further, by taking a large time constant of the loop, the response of the high frequency component of the fluctuation of the frequency of the writing system horizontal synchronizing signal WHD is not responded, and the frequency of the frequency is almost equal to 910 times the average value of the frequency of the writing system horizontal synchronizing signal WHD. The signal can be obtained from the voltage controlled oscillator 191.

Further, in FIG. 31, reference numeral 184 denotes a 4fsc generation circuit for stably generating a signal having a frequency four times as high as the frequency fsc of the NTSC standard color subcarrier with a crystal precision, 185.
Is a read system clock selection switch, and is the second PLL 18
Either the output signal of 3 or the output signal of the 4fsc generation circuit 184 is selected and output as the read system clock RCLK.

Clock generation circuit 157 performs different operations depending on whether or not the VTR for reproducing the video signal (hereinafter referred to as "video signal source VTR") accepts external synchronization. The video signal source VTR is for professional use such as Umatic standard,
The read system clock selection switch 185 selects the output of the 4fsc generation circuit 184 if it has an external synchronizing input, and if the video signal source VTR does not have an external synchronizing input for home use. Select the output of the second PLL 183. This is the field memory 164,165,166
This is because the average frequency of the write clock and the average frequency of the read clock are matched.

First, the operation when the video signal source VTR has an external synchronization terminal (hereinafter referred to as "synchronization mode") will be described. In the synchronous mode, the read system clock selection switch 185 selects the output of the 4fsc oscillation circuit 184. 4fsc oscillator circuit 184 is NT with crystal precision
A signal having a frequency four times the color subcarrier frequency fsc of the SC standard is generated, and is output as the read system clock RCLK via the read system clock selection switch 185. The sync signal generation circuit 158 uses the read clock RCLK as a reference to read the sync signal RSYN.
C is generated and output as an external synchronizing signal of the video signal source VTR.

Since the video signal source VTR performs the reproduction operation of the video signal with reference to the read system synchronizing signal RSYNC, the horizontal frequency of the video signal output from the video signal source VTR, that is, the average frequency of the writing system horizontal synchronizing signal WHD. Becomes equal to 1/910 of the read system clock RCLK. Therefore, in the first PLL 182, the write system horizontal synchronization signal HD is multiplied by 910 to write system clock WC.
Since LK is generated, write system clock WCLK
Changes in accordance with the change in the frequency of the write-system horizontal synchronizing signal WHD, but its average value is equal to the frequency of the read-system clock RCLK.

Next, when the video signal source VTR does not have an external sync input terminal (hereinafter, "stand-alone mode").
Operation) will be described. In the stand-alone mode, the read system clock selection switch 185 selects the output signal of the second PLL 183. First PLL 18
The average frequency of the output signals of both 2 and the second PLL 183 is 9% of the average frequency of the write system horizontal synchronizing signal WHD.
10 times, that is, 4 times the average frequency of the color subcarrier frequency fscW of the input video signal, but the second PLL
Since the time constant of the loop of 183 is set large,
The output of PLL 182 is a signal write system horizontal synchronization signal WH.
The frequency of the output signal of the second PLL 183 gradually follows the change of the frequency fHW of the write system horizontal synchronizing signal WHD, while the change of the frequency of D changes rapidly, and the average of the write system clock WCLK is averaged. It is almost equal to the frequency.

By the way, the reproduction video signal of the VTR such as the Umatic standard, the betamax standard, the VHS standard and the 8 mm standard is not corrected for the time base fluctuation of the luminance signal except for some models, but the color signal is not corrected. All time-axis fluctuations of are corrected and output. Therefore, it is necessary to give the color signal the same time axis variation as the luminance signal before performing the time axis correction. Therefore, when the color decoder 160 converts the color signal into the two color difference signals RY and BY, a signal obtained by dividing the write system clock WCLK by 4 by the 4 divider circuit 159 is used as the reference signal. . That is,
The write system clock WCLK is generated by multiplying the horizontal synchronizing signal including the time axis fluctuation by 910 times as described above, and its frequency corresponds to four times the color subcarrier frequency. Therefore, a signal obtained by dividing the write system clock WCLK by 4 corresponds to a color subcarrier including a time base fluctuation, and a color difference signal R- demodulated using this as a reference signal.
Y and BY have the same time-axis fluctuation as the luminance signal.

The two color difference signals RY, BY and the luminance signal output from the color decoder 160 are converted into digital signals at the timing of the write system clock WCLK by the A / D converters 161, 162, 163, respectively. It is written in the field memories 164, 165, 166 and read out after one field period. The write control of the field memories 164, 165, 166 is performed by the write system memory controller 168 based on the write system clock WCLK, and the read control is performed by the read system memory controller 169 based on the read system clock RCLK.

Next, the time axis fluctuation correction operation will be described. The write clock and read clock of the field memories 164, 165, 166 are clock generation circuits, respectively.
It is generated from the write-system clock WCLK and the read-system clock RCLK output by 157, and the average frequency of both is 910 times the horizontal synchronization frequency. This corresponds to recording the video signal for one horizontal period as 910 pixel data.

Further, since the write-system clock WCLK is generated by multiplying the horizontal synchronizing signal of the input video signal having the time axis fluctuation by 910 times, the frequency thereof also changes momentarily according to the time axis fluctuation. Is the read system clock R
The frequency of CLK is almost constant at four times the frequency of the NTSC standard color subcarrier of crystal precision, or 910 times the average value of the horizontal synchronizing frequency of the input video signal. Therefore,
The operation of the field memories 164, 165, 166 is to write to the storage area corresponding to each pixel of the input video signal at a timing matched to the time axis fluctuation, and to read this regularly, and the time axis of the read signal Will be stabilized.

At this time, if the VTR on the reproducing side can be externally synchronized, the read operation of the field memories 164, 165, 166 is performed with reference to the 4fsc clock of crystal precision, so that the output image with higher time axis precision is obtained. You can get a signal. The field memories 164, 165, 166
It goes without saying that the digitized video signal to be written in does not have to be the entire video period, and only the effective video period is required for both the vertical and horizontal directions.

The image shake detecting section 167 detects the motion vector of the entire image for each field by a known representative point matching method. The representative point matching method obtains a motion vector for each field from the inter-field correlation between some pixels selected as representative points and the surrounding pixels. FIG. 34 is a block circuit diagram of an image shake detection unit 167 by a general representative point matching method, and FIG. 35 is a diagram showing a relationship between an image block and a representative point. An image of one field is divided into a predetermined number of blocks 211, and one representative point Rij 212 is provided at the center of each block. For each block, the level difference between the representative point one frame before and all the pixels Pij (x, y) 213 in the block is calculated.

In FIG. 34, 201 is an address controller for controlling the operation timing of the entire circuit, 202 is a first latch circuit for latching the input luminance signal at the pixel timing of a predetermined representative point, and 203 is a first latch circuit. Latch circuit 202
The representative point memory which writes the data of the pixel of the representative point latched in 1. is stored after reading for one field, and is read out. 204 is a second latch circuit which latches the data of the pixel of the representative point read out from the representative point memory 203. Third time to latch the pixel data at the timing of performing the correlation calculation with the representative point
Latch circuit 206, an absolute value calculation circuit 206 for calculating an absolute value of a difference between pixel data output by the second latch circuit 204 and pixel data output by the third latch circuit 205, and 207 a difference For the output of the absolute value calculation circuit 206, a cumulative addition circuit that cumulatively adds the calculation results of pixels that have the same positional relationship with the representative point for each representative point, and 208 refers to the contents of the cumulative addition circuit 207. Is a motion vector determination circuit that determines a motion vector.

Next, the operation will be described by taking as an example the calculation for the pixels in the block 211. A predetermined pixel in the block 211, which should be the representative point 212, is written in a predetermined area of the representative point memory 203 via the first latch circuit 202. The data stored in the representative point memory 203 is 1
The second latch circuit 20 is field-delayed and read out.
It is sent to the absolute value calculation circuit 206 via 4. On the other hand, the data of the video signal of the current field is the third latch circuit 205.
Is sent to the absolute value calculation circuit 206 via.

The representative point signal of one field before output from the second latch circuit 204 and the pixel signal of the current field output from the third latch circuit 205 are the absolute value operation circuit.
At 206, the absolute value of the difference is calculated. These calculations are performed in block units, and this absolute value calculation circuit 20
The output signal h of 6 is sequentially added to the table corresponding to the same address of the pixel in each block of the cumulative addition circuit 207. The addition result of this table is the motion vector determination circuit 20.
It is input to 8 and finally, the motion vector value is determined in which direction the image position has moved in one field with the block address having the minimum value of the addition result.

That is, the absolute value of the difference between the representative point Rij and the signal Sij (x, y) having the positional relationship in the horizontal direction x and the vertical direction y is obtained, and x, which has the same positional relationship with respect to each representative point,
The values of y are added to obtain the cumulative addition table Dxy as in the following equation. Dxy = Σ | Rij-Sij (x, y) | Then, the minimum values x and y in Dxy are set as the horizontal and vertical motion vectors.

The operation of the above circuit elements is controlled by the control signal from the address controller 201. By the way, if the input image includes a time base fluctuation and is operated with a constant reference clock, the relative relationship between the representative point for each field and the surrounding pixels changes, which adversely affects the motion vector detection. In this example,
The write system clock WCLK is used as a reference clock for the operation of the image shake detection unit 167. Since the timing of the writing-system clock WCLK changes according to the time-axis fluctuation of the input video signal, the sampling timing of the pixels also changes, and the positional relationship of the pixels within the screen becomes substantially constant. Therefore, it is possible to remove the influence of time axis fluctuation.

The read-side memory controller 169 receives the field memory 164, in response to the output of the image shake detection unit 167.
The reading positions of 165 and 166 are controlled to move to stabilize the position of the image in the reading frame. This principle is the same as the conventional example.

As shown in FIG. 44, the field memory 16
The 1 field of the image formed by the output signals of 4, 165 and 166 is smaller than the effective screen size of the image of the 1 field formed by the input video signal for the shake correction operation. Therefore, the D / A converters 173, 174, 1 are used after the interpolation enlargement circuits 170, 171, 172 are used to enlarge the original screen size.
It is converted to an analog signal by 75.

Further, the luminance signal component output from the D / A converter 175 is read by the sync signal replacement circuit 176 from the sync signal generation circuit 158 and the read system sync signal RSYN.
C is added to the luminance signal output terminal 179 and the Y / C mixing circuit 180. In addition, the two systems of color difference signals output from the D / A converters 173 and 174 are converted into an NTSC standard chroma signal by the parallel modulation circuit 177 and then output to the chroma signal output terminal 178 and the Y / C mixing circuit 180. Sent to and.
The Y / C mixing circuit 180 mixes the luminance signal output by the sync signal replacement circuit 176 and the chroma signal output by the parallel modulation circuit 177 to obtain a composite video signal and outputs it to the composite video signal output terminal 181. Chroma signal output terminal 178 for Y / C separation output, luminance signal output terminal 17
9 and the composite video signal output terminal 181
It is possible to obtain a video signal in which the fluctuation of the time axis and the shake of the image are corrected.

As described above, in the fifteenth embodiment, the time base error of the input video signal is detected, the phase of the write clock of the memory is adjusted according to the detected output, and this is adjusted to a clock at a constant interval. The time axis of the video signal is corrected by reading with, and at the same time, the shake of the entire image formed by the video signal is detected, and the shake of the entire image is corrected by moving the read position of the field memory according to the detection result. can do.

Since the image shake detecting section is constructed as shown in FIG. 34, the motion vector can be accurately detected without being affected by the time axis fluctuation of the video signal. Further, the read clock for controlling the read phase of the field memory is switched to either a highly accurate internally generated clock or a clock having an average frequency of the write clock of the field memory depending on whether or not the reproducing VTR can be externally synchronized. Therefore, it is possible to prevent the write operation and the read operation in the field memory from conflicting with each other.

(Sixteenth Embodiment) Next, a sixteenth embodiment in the case where the image shake correcting apparatus is incorporated in the VTR will be described.
FIG. 36 is a block diagram showing the overall configuration of this 16th embodiment,
The same reference numerals as those in FIG. 30 denote the same parts. Figure 36
221 is an electromagnetic conversion unit that records an input electric signal as magnetic information on a magnetic tape and reproduces magnetic information into an electric signal, and 222 is a recording unit that selects either a recording signal or a reproduction signal and outputs it to the electromagnetic conversion unit 221. This is a reproduction signal changeover switch.

Reference numeral 223 denotes a reproduction signal processing unit for processing the output signal of the electromagnetic conversion unit 221 to obtain an NTSC standard video signal, and 224 for processing an NTSC standard video signal to obtain a VHS format recording signal and recording. Playback signal selector switch
A recording signal processing unit for outputting to the electromagnetic conversion unit 221 via 222, a reproducing signal processing unit 223 and an input changeover switch 225.
And 2 are correction input signal selection switches for selecting and outputting either one of the video signals to be output.

Further, 226 is a correction input signal selection switch 22.
The time axis fluctuation / image shake correction unit that corrects the time axis fluctuation of the video and the image shake selected in 5
The servo circuit that controls the electromagnetic conversion unit 221 based on the reference synchronization signal output by the image shake correction unit 226, and 228 selects either the output signal of the time axis fluctuation / image shake correction unit 226 or the input changeover switch 155. And an output signal selection switch for outputting to the recording signal processing unit 224.

FIG. 37 is a block circuit diagram showing the configuration of the time axis fluctuation / image shake correction section 226. In FIG. 37, FIG.
The same reference numerals are given to the same portions as, and the description thereof will be omitted because these portions perform exactly the same operation.

Next, the operation will be described. The system shown in FIG. 36 has two operation modes, a reproduction mode and a recording mode. First, the reproduction mode will be described.
In the reproduction mode, the recording / reproduction signal changeover switch 222 connects the electromagnetic conversion unit 221 and the reproduction signal processing unit 223, and the correction input signal selection switch 225 connects the reproduction signal processing unit 223 and the time axis fluctuation / image shake correction unit 226. Connecting. Further, as the read system clock RCLK output by the clock generation circuit 157, a clock having a crystal precision frequency of 4 fsc is output as in the case where the reproducing side VTR described in the fifteenth embodiment receives external synchronization.

The signal read from the magnetic tape of the electromagnetic converter 221 is converted into a luminance signal and a chroma signal by the reproduction signal processor 223. The output of the reproduction signal processing unit 223 is sent to the time axis fluctuation / image shake correction unit 226 via the correction input signal selection switch 225. The time axis fluctuation / image shake correction unit 226 corrects the time axis fluctuation and the image shake of the video signal output from the correction input signal selection switch 225, generates a read system synchronization signal RSYNC, and outputs it to the servo circuit 227. . At this time, since the reading operation of the field memories 164, 165, 166 in the time axis fluctuation / image shake correction unit 226 is performed with reference to the crystal precision 4fsc clock, it is possible to obtain an output video signal with high time axis accuracy. You can

The servo circuit 227 uses the read system synchronization signal RS.
The reproduction operation of the electromagnetic conversion unit 221 is controlled according to YNC.
Since the read system synchronization signal RSYNC is generated based on the read clock RCLK of the field memories 164, 165, 166 in the time axis fluctuation / image shake correction unit 226, the reproduction operation of the electromagnetic conversion unit 221 and the field memory 164, 165,
It is performed in synchronization with the read operation of 166. The video signal output by the time axis fluctuation / image shake correction unit 226 is sent to the output terminals 178 and 179 and the Y / C mixing circuit 180 via the output signal selection switch 228. The Y / C mixing circuit 180 mixes the luminance signal output from the output selection switch 228 and the chroma signal and outputs the mixed signal to the output terminal 181.

In the playback mode, it is possible to correct the time base fluctuation of the video signal played back by itself and the shake of the image as described above, and to output the stabilized video signal to the monitor or the recording VTR. it can.

Next, the recording mode will be described. If the VTR that is the signal source is a general household VTR,
Normally, the reproduced video signal is not time-axis corrected. However, by using the apparatus of the sixteenth embodiment for a recorder at the time of editing, it is possible to remove the time base fluctuation of the recorded image in advance.

In the recording mode, the recording / reproducing signal selection switch 222 includes the electromagnetic conversion unit 221 and the recording signal processing unit 224, and the correction input selection switch 225 includes the input changeover switch 155 and the time axis fluctuation / image shake correction unit 226. Connecting. Further, as the read system clock RCLK output from the clock generation circuit 157, the reproduction side VTR described in the fifteenth embodiment is used.
Write clock W
A clock having an average frequency of the frequencies of CLK is output. The video signal from the VTR, which is the video signal source, is input to the input terminals 151, 152 or the input terminal 153. One of the chroma signal and the luminance signal input to the input terminals 151 and 152 or the luminance signal and the chroma signal separated by the Y / C separation circuit 154 is selected by the input selector switch 155.

The chroma signal and the luminance signal selected by the input changeover switch 155 are sent to the time axis fluctuation / image shake correction unit 226 via the correction input signal selection switch 225,
The time axis fluctuation component and the image shake are corrected. The output signal of the time axis fluctuation / image shake correction unit 226 is sent to the recording signal processing unit 224 via the output signal selection switch 228, and V
It is converted into a recording signal of HS standard. Recording signal processing unit 224
Output signal is sent to the electromagnetic converter 221 and recorded on the magnetic tape. Further, since the recording operation of the electromagnetic conversion unit 221 is controlled by the servo circuit 227 with reference to the read system synchronization signal RSYNC output from the time-axis fluctuation / image shake correction unit 226, the time-axis fluctuation / image shake correction unit 226's There is no conflict between the reading operation of the field memory and the recording operation of the electromagnetic conversion unit 221.

In the recording mode, it is possible to correct and record the time-axis fluctuation of the video signal input from the outside and the shake of the image as described above.

In the fifteenth and sixteenth embodiments described above, the read system clock WCLK is generated with the synchronizing signal of the input video signal as a reference, but other reference signals such as a burst signal may be used. Further, although the motion vector detecting device by the representative point matching method is used for the image shake detecting unit, the present invention is not limited to this, and another motion vector detecting device such as a gradient method or a Fourier transform method may be used. Although the video signal is set to the NTSC standard, it is not limited to this, and P
Other types such as AL and SECAM may be used. Also,
Although the video signal recording / reproducing apparatus is the VHS standard VTR, the video signal recording / reproducing apparatus is not limited to the VHS standard VTR and may be another system such as a VTR of another standard such as β standard or 8 mm standard, or an optical disk. Furthermore, the video signal is a luminance signal and a color difference signal (RY, BY).
However, it is not limited to this.
Processing may be performed in a B, Y / C separate, Y / C composite, or other state.

Therefore, in the sixteenth embodiment, by controlling the reproduction operation of the VTR with reference to the read clock output from the clock generation circuit, the phase of the reproduction operation of the VTR and the phase of the read operation of the field memory are synchronized. It is possible to obtain an output video signal with time-axis accuracy based on a highly accurate internal clock.

[0207]

As described above, according to the first aspect of the present invention, a stable image can be obtained even when the subject moves, and the subject can be photographed with a fixed composition only by pointing the camera in the approximate direction in which the subject may be present. can do.

In the second invention, as in the first invention, a stable image can be obtained even when the subject moves, and
You can check the movement of the subject.

In the third invention, similar to the first invention, a stable image can be obtained even when the subject moves, and
A stable image can be obtained for an arbitrary subject.

In the fourth invention, as in the first invention, a stable image can be obtained even when the subject moves, and
An image without image quality deterioration can be obtained.

In the fifth invention, as in the first invention, a stable image can be obtained even when the subject moves, and
You can always get a focused image.

According to the sixth aspect of the invention, a sufficient camera shake correction capability can be obtained even in a control system having a relatively low resolution. Further, since the resolution of the control system can be set low with respect to the required camera shake correction ability, the hardware scale can be reduced. Also, since the control system used for camera shake correction is shared, it is possible to prevent an increase in hardware scale and to perform high-speed mode switching.

According to the seventh and tenth inventions, the still state of the photographing device can be detected with certainty.

According to the eighth and ninth aspects, it is possible to reduce the discomfort felt by the operator when switching the modes.

In the eleventh invention, power consumption can be reduced.

In the twelfth invention, since the detection area can be moved, for example, the user specifies the detection area in consideration of the pattern, the position of the subject, etc., and operates so as to eliminate the blurring of that portion. be able to.

In the thirteenth invention, since the detection area is moved according to the judgment result of the reliability of the motion vector detected in each area, the detection area is automatically set to the ideal pattern or subject without the user's designation. Can be set.

In the fourteenth invention, since the movable area is moved so as not to use the same detection point when it overlaps with the fixed area, the same detection point is not detected redundantly but a larger number is detected. It can be detected from the detection point, and the detection accuracy is improved.

In the fifteenth aspect of the invention, the tracking area in the tracking mode and the movable area at the time of shake correction can be combined so that both functions can be realized without increasing the circuit scale.

In the sixteenth invention, it is possible to realize a blur correction device equipped with a motion vector detecting device having the effects of the twelfth to fifteenth inventions.

In the seventeenth aspect of the invention, the time axis error of the input video signal is detected, the phase of the write clock of the memory is adjusted according to the detected output, and this is read out at the clock of a constant interval to read the video signal. At the same time as correcting the time axis of (1), the shake of the entire image can be corrected by moving the read position of the memory according to the detection result of the shake of the entire image formed by the video signal.

In the eighteenth invention, one field or one field is defined with a plurality of specific pixels among the pixels forming the image as a representative point.
Since the data is stored between frames, the correlation between the representative point and the surrounding pixels in one field or one frame is detected, and the writing phase and the phase of the correlation detection operation when storing the representative point are controlled according to the writing clock. The motion vector can be accurately detected without being affected by the time-axis fluctuation of the video signal.

In the nineteenth invention, the read clock for controlling the read phase of the memory is either a highly accurate internally generated clock or a clock having an average frequency of the write clock of the memory, depending on whether or not the reproducing VTR can be externally synchronized. Since the switching is performed, it is possible to prevent the write operation and the read operation in the memory from conflicting with each other.

In the twentieth aspect of the invention, the image shake correction device is a VTR.
Since it is incorporated in the VTR to control the reproducing operation of the VTR with reference to the read clock output from the clock generating means, it is possible to achieve high precision while synchronizing the phase of the reproducing operation of the VTR and the phase of the memory reading operation. It is possible to obtain an output video signal having time-axis accuracy with respect to the internal clock of.

[Brief description of drawings]

FIG. 1 is a block diagram of an image pickup apparatus according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a face image recognition circuit.

FIG. 3 is a diagram showing a range of colors detected as a skin color.

FIG. 4 is a diagram showing how the image pickup apparatus according to the first embodiment operates.

FIG. 5 is a block diagram of an image pickup apparatus according to a second embodiment of the present invention.

FIG. 6 is a block diagram of an image pickup apparatus according to a third embodiment of the present invention.

FIG. 7 is a diagram showing how a subject is recognized in the third embodiment.

FIG. 8 is a diagram showing a range of colors to be detected.

FIG. 9 is a block diagram showing an image pickup apparatus according to a fourth embodiment of the present invention.

FIG. 10 is a block diagram showing an image pickup apparatus according to a fifth embodiment of the present invention.

FIG. 11 is a configuration diagram of an image pickup apparatus according to a sixth embodiment of the present invention.

FIG. 12 is an explanatory diagram of the mechanical system of FIG. 11.

13 is a block diagram of the control system of FIG.

FIG. 14 is a flowchart illustrating an algorithm when switching modes in the sixth embodiment.

FIG. 15 is a graph showing an example of a change in gain of the angle loop in the sixth embodiment.

FIG. 16 is a graph showing another example of the gain change of the angle loop in the sixth example.

FIG. 17 is a configuration diagram showing an image pickup apparatus according to a seventh embodiment of the present invention.

FIG. 18 is a flowchart illustrating an algorithm when switching modes in the seventh embodiment.

FIG. 19 is a configuration diagram of an image pickup apparatus according to an eighth embodiment of the present invention.

FIG. 20 is a flowchart illustrating an algorithm when switching modes in the eighth embodiment.

FIG. 21 is a configuration diagram of an image pickup apparatus according to a ninth embodiment of the present invention.

FIG. 22 is a configuration diagram of an image pickup apparatus according to a tenth embodiment of the present invention.

FIG. 23 is a configuration diagram of an image pickup apparatus according to an eleventh embodiment of the present invention.

FIG. 24 is a block diagram of a motion vector detection device according to a twelfth embodiment of the present invention.

FIG. 25 is a block diagram of an offset setting circuit in a twelfth embodiment.

FIG. 26 is a diagram showing an example of a shooting screen according to the twelfth embodiment.

FIG. 27 is a diagram showing an example in which detection blocks of a fixed area and a movable area are prevented from overlapping in the thirteenth embodiment of the present invention.

FIG. 28 is a diagram showing an example of a CT area movement algorithm in the thirteenth embodiment.

FIG. 29 is a block diagram of an image shake correction device according to a fourteenth embodiment of the present invention.

FIG. 30 is a block diagram of an image shake correction device according to a fifteenth embodiment of the present invention.

FIG. 31 is a block diagram of a clock generation circuit in the fifteenth embodiment.

FIG. 32 is a block diagram of a first PLL in the clock generation circuit in the fifteenth embodiment.

FIG. 33 is a block diagram of a second PLL in the clock generation circuit in the fifteenth embodiment.

FIG. 34 is a block diagram of an image shake detecting section in the fifteenth embodiment.

FIG. 35 is a diagram showing the relationship between image blocks and their representative points for explaining the data processing method according to the fifteenth embodiment.

FIG. 36 is a block diagram of an image shake correction device according to a sixteenth embodiment of the present invention.

FIG. 37 is a block diagram of a time axis correction / image shake correction unit in the sixteenth embodiment.

FIG. 38 is a block diagram of a conventional imaging device.

FIG. 39 is a configuration diagram of a conventional imaging device.

FIG. 40 is a block diagram of a conventional motion vector detection device.

FIG. 41 is a diagram showing a detection area in a conventional motion vector detection device.

FIG. 42 is a diagram showing typical four types of patterns of correlation values (cumulative addition values) with respect to displacement.

FIG. 43 is a block diagram of a conventional image shake correction apparatus.

FIG. 44 is a diagram for explaining the principle of image shake correction.

FIG. 45 is a diagram showing a conventional image shake correction apparatus equipped with a camera-integrated VT.
It is a block diagram which shows the structure at the time of incorporating in R.

[Explanation of symbols]

 1 lens barrel 2 CCD element (light receiving surface) 3 camera signal processing circuit 4 focus detection circuit 5 focus lens control circuit 6 motor driver 8 focus lens position detection device 9 zoom lens control circuit 10 motor driver 12 zoom lens position detection device 14 face image Recognition circuit 15 Zoom control circuit 16 Electronic zoom circuit 17 EVF image control circuit 18 EVF 19 Subject recognition circuit 20 Subject registration circuit 21 Zoom lens 22 Focus lens 30 Lens barrel control circuit 51 Lens barrel 52 Gimbal mechanism 53 Actuator 54 Angle sensor 55 Angular velocity sensor 58 Stationary state detection circuit 59 Adder 60 Actuator drive circuit 61 Image sensor 62 Signal processing circuit 63 Display device 91 Sensor 92 Tripod 93 Housing 94 Sensor 96 Fixed device 101 Representative point memory 102 Absolute value calculation 115 CT area Gate circuit 116 CT area Cumulative addition circuit 117 CT area reliability judgment Path 118 Motion vector determination circuit 119 CT area offset setting circuit 137 Video memory 138 Motion vector detection device 139 Control circuit 156 Sync separation circuit 157 Clock generation circuit 158 Sync signal generation circuit 164 Field memory (RY) 165 Field memory (B- Y) 166 Field memory (Y) 167 Image shake detection unit 168 Write memory controller 169 Read memory controller 182 First phase-locked loop (first PLL) 183 Second phase-locked loop (second PLL) 184 4fsc generation circuit 185 Read system clock selection switch 186 Phase comparator 187 Voltage controlled oscillator 188 910 Frequency divider circuit 189 Phase comparator 190 Low pass filter (LPF) 191 Voltage controlled oscillator 192 910 Frequency divider circuit 201 Address controller 203 Representative point memory 206 Absolute Value calculation circuit 207 Cumulative addition circuit 208 Movement Vector determination circuit 221 Electromagnetic conversion unit 223 Playback signal processing unit 224 Recording signal processing unit 226 Time axis fluctuation / image shake correction unit

Continuation of the front page (31) Priority claim number Japanese Patent Application No. 4-318485 (32) Priority Day Hei 4 (1992) November 27 (33) Country of priority claim Japan (JP) (72) Inventor Shigeki Tsuji Kyoto No. 1 Baba Institute, Nagaokakyo City Mitsubishi Electric Corporation

Claims (20)

[Claims] 1. An optical zoom means for optically zooming, an electronic zoom means for enlarging a digital video signal, a face image recognition means for recognizing a human face portion from the video signal, and a position of a face area. An image pickup apparatus comprising: a zoom control unit that controls the optical zoom unit and the electronic zoom unit based on information on the size and the size. 2. An optical zoom means for optically zooming, an electronic zoom means for enlarging a digital video signal, a face image recognition means for recognizing a human face portion from the video signal, and a position of a face area. And zoom control means for controlling the optical zoom means and the electronic zoom means on the basis of the information on the size and size, and display means for displaying an image before the image is enlarged by the electronic zoom means. Image pickup device. 3. Optical zoom means for optically zooming, electronic zoom means for enlarging a digital video signal, subject recognition means for recognizing a registered subject from the video signal, and position and size of the recognized subject. An image pickup apparatus comprising: a zoom control unit that controls the optical zoom unit and the electronic zoom unit based on information of 1. 4. An optical zoom means for optically zooming, a face image recognition means for recognizing a human face portion from a video signal, an optical axis moving means for moving an optical axis, and a position of a face area. An image pickup apparatus comprising: a zoom control unit that controls the optical zoom unit and the optical axis moving unit based on size information. 5. An optical zoom means for optically zooming, an electronic zoom means for enlarging a digital video signal, a face image recognition means for recognizing a human face portion from the video signal, and a position of a face area. And a size, the zoom control means for controlling the optical zoom means and the electronic zoom means, and the information on the position and size of the face area are used to change the detection area for detecting the focus position. An image pickup apparatus comprising: a changing unit. 6. A lens barrel having an optical system for photographing a subject, photoelectric conversion means for converting an optical image from the lens barrel into an electric signal, and holding means for holding the lens barrel and the photoelectric conversion means. Rotating support means for rotatably supporting the holding means in the panning direction and / or the pitching direction, actuator means for rotatably driving the rotating support means, and detecting a relative angle between the holding means and the apparatus housing. Detecting means, return means for controlling the actuator means by a control value proportional to the relative angle detected by the detecting means to return the holding means to the reference position, and stationary detection for detecting the stationary state of the housing. And an image pickup device configured to increase a control gain to the returning unit when the stationary state detecting unit determines that the housing is stationary. 7. The stationary state detecting means determines that the stationary state is a stationary state when the angular velocity of the housing is smaller than a predetermined value for a certain period of time, and a normal state when the angular velocity of the housing is a predetermined value or more. The image pickup apparatus according to claim 6, wherein the image pickup apparatus is configured as described above. 8. The image pickup apparatus according to claim 6, wherein when the stationary state continues, the control gain for the relative angle is increased with the elapse of the stationary time. 9. The image pickup apparatus according to claim 6, further comprising display means for displaying the static state to the operator when the static state is detected. 10. The image pickup apparatus according to claim 6, wherein the stationary state detecting unit is configured to determine that the stationary state is set when a fixing device such as a tripod is attached to the housing. 11. A lens barrel having an optical system for photographing a subject, photoelectric conversion means for converting an optical image from the lens barrel into an electric signal, and holding means for holding the lens barrel and the photoelectric conversion means. Rotating support means for rotatably supporting the holding means in the panning direction and / or the pitching direction, actuator means for rotatably driving the rotating support means, and detecting a relative angle between the holding means and the apparatus housing. Detecting means, a first returning means for returning the holding means to the reference position by controlling the actuator means with a control value proportional to the relative angle detected by the detecting means, and detecting the stationary state of the housing. And a second returning means for mechanically fixing the holding means. When the stationary detection means determines that the housing is stationary, the second returning means is operated. Movement An image pickup device, characterized in that the holding means is configured to return to a reference position. 12. A means for obtaining a correlation value at a predetermined deviation between screens for each detection area by providing a plurality of detection areas in the screen, and a means for obtaining a motion vector for each detection area from the correlation value. A motion vector detecting device comprising: a means for moving a predetermined number of detection areas among a plurality of detection areas; and a means for determining a motion vector of the entire screen using the motion vector of each area. 13. A means for determining a correlation value in a predetermined deviation between screens for each detection area by providing a plurality of detection areas in the screen, and a means for determining a motion vector for each detection area from the correlation value. A means for determining the reliability of the motion vector detected from each detection area, a means for moving a predetermined number of detection areas among the plurality of detection areas, and a motion vector for each area based on the reliability determination result. A motion vector detecting device, comprising means for determining a motion vector of the entire screen by using it and moving a movable detection area. 14. A means for determining a correlation value at a predetermined deviation between screens for each detection area by providing a plurality of detection areas in the screen, and a means for determining a motion vector for each detection area from the correlation value. A correlation value detection point of the movable detection area is provided, which comprises means for moving a predetermined number of detection areas among the plurality of detection areas and means for determining the motion vector of the entire screen using the motion vector of each area. Is configured so that it does not overlap with correlation value detection points of other detection areas. 15. A means for determining a correlation value at a predetermined deviation between screens for each detection area by providing a plurality of detection areas in the screen, and a means for calculating a motion vector for each detection area from the correlation value. A means for moving a predetermined number of detection areas among the plurality of detection areas and a means for determining the motion vector of the entire screen using the motion vector of each area are provided. A motion vector detecting device having a tracking mode for moving in accordance with a detected motion vector. 16. The motion vector detecting device according to claim 12, 13, 14 or 15, storage means for storing a video signal of one field or more, and a motion vector detected by the motion vector detecting device. In addition, the image shake correction apparatus is provided with a control unit that controls a read position of the storage unit. 17. A time axis fluctuation detecting means for detecting a time axis fluctuation of a video signal, a memory capable of writing and reading the video signal, an image shake detecting means for detecting a shake of an image from the video signal, and the time axis fluctuation. A write clock that controls the write phase of the video signal of the memory and a read clock that controls the read phase of the memory at a stable cycle are generated so as to correct the time base fluctuation of the video signal according to the detection output of the detection means. Clock generation means, memory write control means for controlling the write operation of the memory according to the timing of the write clock output by the clock generation means, and memory write control means for the memory according to the timing of the read clock output by the clock generation means. Depending on the read-out phase and the amount of shake detected by the image shake detecting means, An image shake correction apparatus, comprising: a memory read control unit that reduces image shake by controlling each read address of the memory. 18. The image shake detection means, a representative point storage means for storing a plurality of specific pixels among pixels forming an image as a representative point for one field or one frame, the representative point and its surroundings. Correlation detection means for detecting the correlation between one pixel and one field or one frame, and the write phase of the representative point storage means and the correlation detection operation of the correlation detection means according to the write clock generated by the clock generation means. 18. The image shake correction apparatus according to claim 17, further comprising a control unit that controls a phase. 19. The clock generating means, write clock generating means for generating the write clock, first read clock generating means for generating a clock having a frequency corresponding to an average frequency of the write clock, and crystal precision. Second to generate a constant frequency clock of
Read clock generating means, read clock selecting means for selecting and outputting either the output of the first read clock generating means or the output of the second read clock generating means, and the second read clock generating means. 18. The image shake correction apparatus according to claim 17, further comprising: a reading system synchronizing signal generating means for generating a synchronizing signal based on the output signal of the means. 20. A video signal recording / reproducing means is included, and a reproducing operation of the video signal recording / reproducing means is controlled on the basis of a read clock output from the clock generating means. The image shake correction device described.