JPH0767024A - Compound eye type image pickup device - Google Patents

Abstract

PURPOSE:To easily designate and select an object for an observer while performing photographing by controlling the turning angle of one of first and second video cameras at least. CONSTITUTION:When an initializing mode and a stereoscopic photographing mode are selected, the zoom optical system of two camera parts A and B are set to an initial value first, and an interval between two camera parts A and B is initialized later. Further, the camera angle is initialized so that two camera parts A and B can be made parallel. Next, photographing is performed while measuring a distance from a distance detector 31 of the camera part selected by a main camera selecting SW 25 such as the right camera part A to an object in front of the camera part A and the camera angle. Next, when the initializing mode and a panoramic photographing mode are selected, a control part 15 measures a distance to overlap the image angles of two cameras, and two cameras are focused to perform photographing based on that value.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compound eye image pickup device using two video cameras.

[0002]

2. Description of the Related Art Conventionally, as a compound-eye image pickup apparatus of this type, there is a stereoscopic image pickup video camera disclosed in Japanese Patent Publication No. 5-3199. This camera body has two video cameras oriented substantially in the same direction. A stereoscopic image is obtained from the images obtained by shooting the subject directly facing the camera from different angles. Further, in the stereoscopic image pickup focusing device disclosed in Japanese Patent Laid-Open No. 62-12493, in order to photograph an object that does not face the camera body, an infrared spot light is emitted to the object to be photographed and the reflected light is reflected. Orienting the optical axis of the lens is disclosed.

[0003]

However, of the above-mentioned conventional examples, there is a problem that the former cannot specify a subject to be projected. In the latter case, the mechanism for designating the subject is manual, and it is very difficult for the observer to designate the subject while capturing an image.

The present invention has been made in view of the problems of the above-mentioned conventional techniques, and an observer can easily photograph an object not only in the stereoscopic photographing mode but also in the panoramic photographing mode. An object is to realize a compound-eye imaging device that can be designated and selected.

[0005]

In the compound-eye image pickup apparatus of the present invention, the first video camera section and the first video camera section in which the mechanism for setting the image pickup condition of each aperture, focus and zoom is controlled by at least one camera control means. A second video camera unit, a unit for combining the first video camera unit and the second video camera unit, and a plane obtained by each optical axis of the first video camera unit and the second video camera unit. Optical axis direction varying means for varying the optical axis directions of the first and second video camera sections, and the first video camera section and the second video camera section in a plane that is substantially horizontal with respect to An optical axis angle control means for controlling the optical axis angle of the, and a viewfinder for displaying an image captured by at least one of the first video camera section and the second video camera section. In a compound-eye image pickup device provided in the viewfinder and having a subject selecting unit capable of selecting an arbitrary subject in the viewfinder screen, a first video camera unit and a first video camera unit corresponding to an output of the subject selecting unit. The optical axis angle of at least one of the two video camera units is controlled.

[0006]

According to the output of the subject selection means, the first
Since the configuration is such that the rotation angle of at least one of the second video cameras is controlled, the observer can easily designate and select a subject while taking a picture.

[0007]

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

FIG. 1 is a block diagram showing the construction of an embodiment of a compound eye image pickup apparatus according to the present invention, and FIG. 2 is an external perspective view of an embodiment of the present invention.

The first video camera section A and the second video camera section B are arranged substantially in parallel in substantially the same direction and are integrated by a uniting means C. The first and second video camera sections A and B are composed of a main body section, an optical section, an electric circuit section inside the main body, etc., and a semiconductor element using a CCD is used as an image pickup element. In addition, the main body sends the image pickup output corresponding to each video camera to the corresponding storage device such as a VTR, and receives the power supply from the outside,
Furthermore, connector parts 50 and 51 for supplying a remote control signal are provided. Further, in this compound eye image pickup device, a viewfinder 24 for displaying the image pickup output of one of the video cameras is provided in the uniting means C.

By the way, in order to stereoscopically view the images picked up by the first and second video camera units A and B, it is necessary to superimpose the images from both on the screen or the screen of the cathode ray tube. And it is necessary to obtain a three-dimensional effect when they are superposed. For that purpose, it is necessary to cross the optical axes of the first and second video camera units A and B on the object side to capture an image. Here, the relationship between the intersecting position of the optical axes of both cameras and the position of the subject has a great influence on the stereoscopic effect. For example, if the intersecting position and the subject are the same, the position where the stereoscopic image is expressed is almost the same as the screen surface.
However, if the intersecting position is in front of the subject, the three-dimensional representation position is behind the screen. On the contrary, if the intersection position is behind the subject, the three-dimensional expression position is in front of the screen,
It will be in a position protruding from the screen.

In order to properly obtain such a stereoscopic effect, it is necessary to adjust the intersection position of the optical axes of the first and second video camera units A and B according to the distance between the subject and the camera. . The compound-eye imaging device according to the present invention is also configured with this point in mind. Further, in the compound-eye imaging device according to the present invention, not only stereoscopic shooting but also panoramic shooting is possible by switching the SW as described later. Next, the configuration of the compound-eye imaging camera according to the present invention will be specifically described with reference to FIG.

In FIG. 2, the light for the right eye is the lens group 1
The image is picked up by the image pickup element 2 inside the right video camera section A through and output as a right-eye video signal. On the other hand, the light for the left eye is captured by the image sensor 5 inside the left video camera unit B through the lens group 4 and output as a left eye video signal. Each of the right-eye video signal and the left-eye video signal is output to output terminals 50 and 51 through the processing described below.

Motors for driving the respective lens groups 1 and 4 are provided for the left and right cameras respectively for zooming and focusing, and a right zoom motor 7, a right focusing motor 8, a left zoom motor 9 and a left focusing motor are provided. A focus motor 10 is provided. Further, each of the lens groups 1 and 4 is provided with diaphragms 3 and 6 for adjusting the amount of light received by the image sensor. Further, position detectors 11, 12, 13, and 14 for detecting the lens positions driven by the zoom motor and the focus motor are provided, and the outputs thereof are input to the control unit 15. The first and second video camera units A and B are provided with distance detectors 33 and 34, which are composed of, for example, a light emitting element and a light receiving element, and measure the distance to the subject and input it to the control unit 15. . At least one of the distance detectors 33 and 34 is configured to be rotatable in the horizontal direction, and it is possible to measure the distance of the subject in the direction designated by the command of the control unit 15. There is. Reference numeral 16 denotes an EE that stores the variation in the sensitivity of the image sensor, the individual difference between the left and right lens groups, and the initial values of the angles and intervals of the two cameras described later at the time of manufacturing.
It is a PROM.

Further, a right convergence angle motor 17 and a left convergence angle motor 18 for rotating the camera in the horizontal direction are provided with the light receiving surfaces of the image pickup devices 2 and 5 as substantially rotation centers, and the left and right cameras rotate. Angle detectors 19 and 20 are provided, and the outputs thereof are input to the control unit 15. And
The right video camera unit A has an adjusting unit 21 for adjusting the distance between the left and right cameras and an adjusting motor 2 for driving the adjusting unit 21.
2 and a gap detector 23 that detects the gap between the left and right cameras, and the output thereof is input to the control unit 15.

The image pickup devices 2 and 5 are driven in synchronization by a synchronizing signal generator 36, and an analog video signal output from the device is converted into a digital video signal through an amplifier (not shown) and an A / D converter 40. , 41, respectively.
The digital video signals output from the A / D converters 40 and 41 are stored in the image memories 41 and 44 whose reading and reading are controlled by the synchronizing signal generator 36, respectively, and each image is processed by the image correlation processing unit 46. Memory 41, 44
The image correlation calculation is performed from the digital video signal stored in. The calculation result of the image correlation processing unit 46 is input to the control unit 15, and the address read from the memory or the time axis is changed according to the calculation result. Reference numeral 47 is an image composition processing unit that is composed from the calculation results obtained by the image correlation processing unit 44 at the time of panoramic photography, and 42 and 45 are analog signals obtained from the image memories 41 and 44 or the image composition processing unit 47. The signal output to the D / A converter 45, which is a D / A converter for converting into a video signal, has its input switched to either the image memory 44 or the image composition processing unit 47 by the stereoscopic / panorama switching SW 35.

A viewfinder 24 is a D / A converter 4
An analog video signal obtained through 2, 45 can be viewed. Also the viewfinder 24
Is rotatably attached to the uniting means C, and this rotation causes not only the right eye but also the left eye by rotating the eyepiece part of the viewfinder 24 by rotating 180 ° to the side opposite to the camera body. You can see it. Further, although the viewfinder 24 is turned upside down by rotating it by 180 °, the viewfinder 24 is configured to be rotatable by a mechanism (not shown) so that no trouble occurs.

Furthermore, during stereoscopic photography, the main camera selection switch
25, the signal output to the viewfinder 24 can be selected from either the signal from the right video camera unit A or the signal from the left video camera unit B. That is, when the photographer is looking through the viewfinder 24 with the right eye, the main camera can be the right camera, and when looking through the left eye, the main camera can be the left camera. . During panoramic photography, in this embodiment, by switching to the left camera, the panoramic image synthesized from the signals captured by the left and right cameras can be observed by the viewfinder 24. A line-of-sight detector 26 is provided in the viewfinder 24 so that the position of the line of sight can be detected.

Further, as operation buttons or SW on the camera body, a camera standby button 27, a zoom operation button 28, a subject selection button 29, a subject tracking button 30, a status storage button 31, a stereoscopic / panorama switching SW 35, and an initial state. There is a selection SW32,
The initial state selection SW 32 sets the state of the camera in the camera standby state to the initialization state, or stores the state of the camera when the status storage button 31 is pressed, and automatically switches to that state in the standby state. It is something that will be restored.

Next, the zoom focusing system in this embodiment will be described.

In this embodiment, a rear focusing type lens system is adopted, and the function of the correction lens for correcting the movement of the focal plane due to the change of the focal length with respect to the focusing lens 1-a and the function of the focusing lens are as follows. It has both functions. That is, in this type of lens system, since the correction lens also serves as the focusing lens, the movement locus of the correction lens varies depending on the subject distance during zooming. Therefore, the plurality of focusing lens movement loci are stored in the control unit 15, and the loci are selected according to the subject distance to move the focusing lens 1-a.

Further, the difference in the angle of view, the shift of the optical axis at the time of zooming due to the dispersion of the left and right lens groups, and the shift of the timing of focusing are also weak points of the compound-eye image pickup system. The correspondence data of the left and right zoom lens positions and the correspondence data of the distance to the subject and the left and right focusing lens positions at that time are written.

Next, the visual axis detection system in this embodiment will be described.

In the present embodiment, the viewfinder 24
Is equipped with a line-of-sight detector. As a method of detecting the line of sight, for example, in Japanese Patent Application Laid-Open No. 61-172552, a parallel light flux from a light source is projected onto the anterior segment of the eyeball of an observer, and a corneal reflection image and a pupil are combined by reflected light from the cornea. The visual axis is obtained using the image position.

FIG. 3 is a diagram for explaining the principle of line-of-sight detection.
Is a schematic view of the line-of-sight detection optical system, and (b) is an output intensity diagram of the photoelectric element array 66.

In the figure, reference numeral 65 denotes a light source such as a light emitting diode which emits infrared light insensitive to the observer, and is arranged on the focal plane of the light projecting lens 63. The infrared light emitted from the light source 65 is collimated by the light projecting lens 63, reflected by the half mirror 62, and illuminates the cornea 67 of the eyeball. At this time, a corneal reflection image (virtual image) d due to a part of the infrared light reflected on the surface of the cornea 67 passes through the half mirror 62, is condensed by the light receiving lens 64, and is photoelectrical device array 66.
The image is projected at the upper position Zd '. Also, the edge a of the iris 69,
The light flux from b passes through the light receiving lens 64 and the photoelectric element array 6
Images of the end portions a'and b'are formed on a'and b'on 6, respectively.
When the rotation angle θ of the optical axis a of the eyeball with respect to the optical axis (optical axis a) of the light receiving lens 64 is small, assuming that the Z coordinates of the ends a and b of the iris 69 are Za and Zb, the coordinate Zc of the central position c of the pupil is represented.
Is expressed as Zc≈ (Za + Zb) / 2.

Further, since the Z coordinate of the corneal reflection image d and the Z coordinate of the center of curvature O of the cornea 67 coincide with each other, the Z coordinate of the generation position d of the corneal reflection image is Zd, the center of curvature O of the cornea 67 and the pupil. Assuming that the distance to the center C is Oc, the rotation angle θ of the optical axis of the eyeball satisfies the relational expression of Oc × SINθ≈Zc−Zd (1). Therefore, the arithmetic processing unit 60 detects the positions of the respective feature points (corneal reflection image d and the iris ridges a and b) projected on the photoelectric element array 66 as shown in FIG. The rotation angle θ of the optical axis a can be obtained. At this time, the equation (1) can be rewritten as γ × Oc × SINθ≈ (Za ′ + Zb ′) / 2−Zd ′ ... (2). However, γ is a magnification determined by the position of the eyeball with respect to the light receiving lens 64.

Further, when the rotation angle θ of the optical axis of the eyeball of the observer is calculated, the line of sight of the observer can be obtained by correcting the optical axis of the eyeball and the visual axis.

Further, in FIG. 3, the observer's eyeball is Z-.
Although the example of rotating in the X plane (for example, horizontal plane) is shown, it can be similarly detected when the eyeball of the observer rotates in the XY plane (for example, vertical plane).

Next, the operation of the compound-eye image pickup apparatus according to this embodiment will be described.

FIG. 4 is a flow chart showing the control operation by the control unit 15 of this embodiment. The initialization operation of the apparatus of this embodiment will be described with reference to FIG.

When power is supplied to the camera and the camera enters a standby state (step S401), the control unit 15 detects the states of the initial state selection SW 32 and the stereoscopic / panoramic switching SW 35 to determine whether the operation mode of the apparatus is the initialization mode. It is confirmed whether it is the storage mode (step S402). When the initialization mode is selected, the control unit 15 further confirms whether the stereoscopic shooting mode or the panoramic shooting mode (step S).
403).

When the initialization mode and the stereoscopic photographing mode are selected, first, the zoom optical systems of the two cameras are set to the initial values (step S409), and the interval between the two cameras is set to a predetermined initial value. The camera interval for driving the adjustment motor 22 is initialized so that (step S41
0) Further, the left and right camera angles for driving the two cameras so as to be perpendicular to the line connecting the rotation centers of the two cameras and parallel to each other are initialized (step S).
411). These state settings are set by the angle detector 2 at the time of manufacture.
Measured from the reference value of 9 and 20 to the above state,
This is performed based on the angle offset value written in the EEPROM 16.

Next, the camera selected by the main camera selection SW 25, for example, the right first video camera unit A
The distance to the object in front of the right video camera section A and the camera angle θ0 are measured from the distance detector 31 of (4), (steps S412, S413), and an image is taken (step S41).
4).

As shown in FIG. 4, assuming that the distance measurement result is 10 and the camera interval is e0, the camera angle θ0 is θ0 = ATAN (2 × 10 / e0) (Hereinafter, ATAN represents an inverse function of TAN. Then, the control unit 15 directs each camera to the subject, and focusing is performed there. That is, the object is automatically aimed and focused on the vertical bisector of the line connecting the rotation centers of the two cameras.

In the present embodiment, the camera interval is assumed to be about 10 cm at most, so there is no problem even if a subject facing the right video camera unit A is used as the subject on the vertical bisector. There is no. Also, assuming a larger system, for example, even when the distance between the two cameras is several tens of centimeters or several meters, a separate distance sensor may be provided on the vertical bisector. If the detection mode is set such that the light emitting element of the video camera unit A outputs the light from the light receiving element of the left video camera unit B, distance measurement can be performed under substantially the same conditions.
Further, if accuracy is required, it can be realized by arranging the distance detectors of the two cameras symmetrically with respect to the vertical bisector.

Next, the case where the initialization mode and the panoramic photography mode are selected will be described.

As in the case where the initialization mode and the stereoscopic photographing mode are selected, the control unit 15 drives the two cameras so that the two cameras are perpendicular to the line connecting the respective rotation centers and are parallel to each other. Yes (steps S404 to S40)
6). Then, the distance to a portion where the angles of view of the two cameras overlap (again, the subject facing the right camera can be used instead) can be measured (step S407), and the two cameras can be focused based on the value. Let (Step S40
8) Take an image (step S414).

Next, when the storage mode is selected, it is confirmed whether the status at the time of pressing the status storage button 31 is stored in the EEPROM 16 (step S415), and the status storage button 31
When the button is pressed, the state of the stereoscopic / panorama switching SW, the angles, intervals, and zoom positions of the left and right cameras are automatically restored by the reading control unit 15 (step S416) and an image is taken (step S416). S414).

Further, in step S415, the EEPROM is
If it is confirmed that the status is not stored in 16, the process proceeds to step S403, and the initialization mode is automatically set.

Next, the operation of this embodiment at the time of photographing will be described.

In this embodiment, in the stereoscopic photographing mode,
It has the following two line-of-sight detection modes using the above line-of-sight detection. That is, the subject selection mode and the subject tracking mode.

First, the subject selection mode is shown in FIG.
A description will be given using (a) to (d). In the following description, it is assumed that the main camera is the right video camera section A.

FIG. 6A shows a video image being captured in the viewfinder 24. FIG. 6B shows the right video camera unit A.
And the left video camera unit B and the subject are schematically shown. The lines connecting the subjects 70 and 71 and the right video camera unit A are 72 and 73, and the line passing through the subject 70 and perpendicular to the line 72 is 74. To do. In addition, lines 75, 76 connecting the subjects 70, 71 and the left video camera unit B pass through the subject 70, and a line 7
The line perpendicular to 5 is 77.

The subject 70 is located at a position where the optical axes of the left and right cameras intersect with each other, but the subject 71 is located at a position other than that. Appears to be at intersection 78 of 74. On the other hand, when viewed from the left video camera section B, the line 7
Appears to be at intersection 79 of 6 and 77. That is, when these are overlapped, the subject 71 is doubled as shown by the broken line in FIG.

The stereoscopic image obtained at this time is observed such that the subject 70 is on the screen and the subject 71 is on the back side of the screen. Therefore, if the subject selection button 29 is pressed while gazing at the subject 71, the angle of the left video camera unit B is automatically changed so that the images of the subject 71 are overlapped by the left and right cameras. Therefore, if the images captured by the left and right cameras are overlaid at this time, the subject 7
0 is doubled.

This operation will be described with reference to FIG. In FIG. 6C, the same parts as those in FIG.

Right video camera section A and left video camera section B
The line connecting the rotation centers of is 80, and the distance between the centers is e1,
Right video camera section A, left video camera section B, and subject 70
Up to l1 (assuming that the subject 70 is on the vertical bisector of the line 80), the distance between the right video camera unit A and the subject 71 is l2, and the distance between the left video camera unit B and the subject 71 is l3. , The angle between line 80 and lines 72 and 75 is θ1, line 7
The angle between the line 3 and the line 72 is θ2, the angle between the line 76 and the line 80 is θ3, and the line passing through the subject 71 and parallel to the line 74 is 81,
If the line passing through the subject 71 and parallel to the line 77 is 82, the horizontal distance l4 from the screen center 83 of the subject 71 on the screen when observed by the video camera unit A is: l4 = l2 × SINθ2 Can be expressed (because the subject 70 is at the center of the screen). On the other hand, the horizontal distance l5 from the screen center 84 of the subject 71 on the screen when observed by the video camera unit B can be expressed as l5 = l3 × SIN (θ1−θ3). Therefore, in order to superimpose the image of the subject 71 of the left video camera unit B on the image of the subject 71 of the right video camera unit A, the screen of the left video camera unit B may be shifted to the left. The shift amount 16 is 16 = 14-15. To actually perform this, the left video camera unit B may be rotated counterclockwise by θ4 in the figure.
The rotation angle θ4 can be obtained as θ4 = ATAN (16 / (13 × SIN (θ1-θ3)), and by rotating the left video camera unit B by the left convergence angle motor 18 by this angle, the subject 71 It will be in the state of gazing at.

Here, e1 is obtained from the distance detector of the adjusting unit 21, l1 is obtained from the distance detector 33 or the distance detector 34, and θ1 is obtained from the angle detector 19 or the angle detector 20. can get. Further, θ2 is obtained from the line-of-sight detector 26, l2 is obtained by rotating the distance detector by θ2 to measure the distance, and θ3 is obtained from l2, θ1 and θ2.
And e1, θ3 = ATAN (l2 × SIN (θ1 + θ2 / (e1 + l2 × COS (θ1 + θ2)) ... (3) and l3 = l2, θ1, θ2, θ3 Since (θ1 + θ2) / SINθ3 (4) is obtained, the above mode can be easily implemented.

In the actual operation of this mode, the distance detector 33 of the right video camera unit A is obtained from the line-of-sight detector 26 by gazing at the subject to be selected while looking through the viewfinder and pressing the subject selection button. It rotates by θ2, and the distance l2 between the right video camera unit A and the subject 71 is measured. Then, from the data obtained from each detector, that is, e1, l1, l2, θ1, θ2, θ3, l3, θ
4, the focusing data corresponding to 12 and 13 are read from the EEPROM 16 to focus the left and right cameras together, and then the left video camera unit B is rotated based on the obtained θ4.

However, since it is impossible to mechanically eliminate the vertical displacement of the left and right images under all image capturing conditions, the stereoscopic image thus captured contains vertical displacement. And this vertical displacement is one of the factors that hinder stereoscopic viewing. Therefore, the signals captured by the two cameras are once stored in the image memories 41 and 44, and the image correlation processing unit 46 calculates the correlation between the two images. The area to be subjected to correlation calculation is, for example, 3 × 1 on the screen.
It is sufficient to divide into 0 regions and to perform only in the region where the line of sight is detected.

Although there are various correlation calculation methods, the matching method is used in the present embodiment, and the image signals in the same region of the signals picked up by the left and right cameras are binarized and used as the main camera. The binary image obtained from the signal imaged by the selected right video camera unit A is moved by a fixed amount s (for example, one pixel) in the horizontal direction and the vertical direction, and the difference δ from the other binary image is calculated. Ask. Then, the calculation is sequentially shifted to perform the shift amounts X and Y that minimize the difference δ.
Then, the Y value is used to eliminate the vertical shift.

When the two images are shifted by X and Y as described above, the common part of both images is narrower than the imaged area. That is, as shown in FIG. 7, the area p that is effective as an image is a portion excluding the non-common portion q and the portion r to be deleted in order not to change the angle of view. In order to expand this effective area p to all the imaged areas, the cycle of reading the data of the memory by the synchronizing signal generator 36 in the horizontal direction is delayed, and in the vertical direction, it is selected by interpolation and reading. Vertical displacement in the left and right images of the subject disappears, and the stereoscopic image of the selected subject appears to move from the back side of the screen to the screen.

In the above embodiment, the calculation is performed by shifting by one pixel, but for example, the calculation is performed by shifting by 5 pixels, and interpolation is performed in the vicinity of the position where the minimum value of the difference δ is obtained. , Y, the calculation time can be shortened.

Next, the subject tracking mode will be described with reference to FIG. 6 (d). In FIG. 6D, the same parts as those in FIG.

A line 85 passing through the subject 71 and orthogonal to a line 73 connecting the right video camera unit A and the subject 71 is orthogonal to a line 73 passing through the subject 71 and a line 76 connecting the left video camera unit B and the subject 71. 86, an intersection of a line 72 and a line 85 connecting the right video camera unit A and the subject 70 is 87, and an intersection of a line 75 and a line 86 connecting the left video camera unit B and the subject 70 is 88.

When the subject tracking button 30 is pressed in the state of gazing at the subject 71 from the state of FIG. 6B, the distance detector 3
3 rotates clockwise by the angle θ2 obtained by the line-of-sight detector 26 to obtain the distance l2 to the subject 71. Therefore, the angle θ3 for the optical axes of the left video camera unit B and the right video camera unit A to intersect with each other at the subject 71 is calculated from the equation (3),
Then, the positions of the focusing lenses corresponding to l3 and l2 obtained from the equation (4) are read from the EEPROM l6, and both cameras are focused on the subject 71, and then the right video camera unit A and the left video camera unit B are set. Angle θ2, θ3
Rotate clockwise based on. This makes the subject 7
1 automatically moves to the center of the screen. At this time as well, the processing for eliminating the vertical shift is performed as described above and the image is output.

In the present embodiment, the distance measuring method of the type using the light emitting element and the light receiving element as the distance measuring means of the camera has been described. It's a little complicated. Other methods of distance measuring means include phase difference type and TV
There is a signal detection type, and the phase difference type is for focusing the subject based on the amount of displacement of a plurality of subject images formed through different optical paths, but a mechanism for rotating the sensor is not required There is also a problem in that the number of detection units is required for detection. In the TV signal detection type, specifically, a line-of-sight is detected, and a specific high-frequency component (for example, 1 Mhz) of a video signal in a region corresponding to the line-of-sight is extracted by a bandpass filter, and the amplitude of this high-frequency component is maximized. The focus lens is adjusted by adjusting the position of the focusing lens so that the signal processing circuit of the detection system is added and the time until the focus detection is longer than the other two methods. There is also the problem of this.

Each of the above three methods has advantages and disadvantages, but which method is adopted is arbitrary.

Next, the panoramic shooting mode will be described.

During panoramic photography, the two cameras are set to rotate with respect to the vertical bisector of the line connecting the centers of rotation of the two cameras. It is configured to be selectable, and the degree of overlap between the images captured by both cameras is determined by the selected aspect ratio.

FIG. 8 shows an overlapping area in the right video camera section A, and here, the overlapping area can be changed in four steps. By the way, if the distance to the subject is sufficiently large with respect to the camera distance, the ratio κ to the angle of view of the overlapping area of the angle of view of both cameras is α, which is the camera angle from the initial state of the panoramic shooting mode of the camera, and zoom ratio. Let β be a half angle of the horizontal angle of view determined by κ = 1 / (1 + TANα × TAN (α + β)) / (2 × ((1 / (1 ten TAN α × TAN (α-β))-(COSα ) ^ 2)) ······· (5), which means that it is determined by the zoom magnification and the camera angle regardless of the subject distance.Conversely, if the ratio κ is selected and the zoom magnification at that time is known. Since the camera angle can be calculated by the equation (5), for example, when typeII is selected, the camera angle corresponding to the overlapping area is calculated, and the two cameras are rotated to perform the focusing operation in the overlapping area. And The picked-up digital video signals are stored in the image memories 41 and 44, respectively, and the image correlation processing unit 46 performs image correlation calculation from the digital video signals stored in the image memories 41 and 44. Here, the stereoscopic shooting mode is used. Similar to the time, the position of the area to be overlapped at the time of panoramic photography with respect to the screen is measured by the matching method, and the position is stored in the image synthesizing processing unit 47 so as to be overlapped.

The image information stored in the image synthesizing processing unit 47 is larger than the information for one normal screen, and the data in the memory is more than normal in the horizontal direction. In order to speed up the cycle of reading by the generator 36 and to make the horizontal and vertical magnifications the same, it is sufficient to trim the top and bottom of the screen in the vertical direction and read by compressing or thinning.

The output from the image synthesis processing unit 47 is a stereoscopic image.
It is switched by the panorama switching SW 35, and is output to the output terminal 51 via the D / A converter 45. Since the output of the D / A converter 45 is an analog video signal, this signal can be regarded as k which is also observed by the viewfinder 24. Therefore, the aspect ratio can be changed while observing the viewfinder 24. At this time, when the aspect ratio change button is pressed, the camera angle is set, and the screen is configured to output the image stored in the image combining processing unit 47 until the focusing operation is completed. The disordered image inside is not output.

In the above embodiment, the line-of-sight detection is taken as an example of the subject selection method, but the selection method is not limited to this. For example, if the screen is 3 × 10
The area may be divided into areas, and the area may be selected with a button or the like. This area selection means is composed of an area move button and a select button. The area move button is a button for moving the 3 × 10 area displayed in the viewfinder up, down, left and right on the screen, and the select button is the button. The area may be made selectable by pressing.

With the above-described method, the functions of the subject selection mode and the subject tracking mode can be implemented as in the case of utilizing the visual axis detection described above.

The method of selecting the area on the screen with the button is not applied only to the viewfinder, and the remote terminals 50 and 51 of this compound-eye image pickup apparatus are used to detect the subject from an external monitor or the like. It is also possible to make a selection to control the angle of convergence of the camera.

Further, in the above embodiment, the visual axis detection is explained to operate only in the stereoscopic photographing mode, but it can also be used to select the aspect ratio in the panoramic photographing mode. The application method is to combine a region divided into 3 × 10 into a smaller region, for example, a 3 × 2 region into one, divide it into a 1 × 5 region as a whole, and select the region with a line of sight. The angle of the cameras may be changed so that the two cameras can overlap and capture the area.

Although the optical axis can be directed to the subject by rotating the camera in the above embodiment, it is possible to use an optical element such as a variable apex angle prism (hereinafter referred to as VAP). It is possible to direct the optical axis to the subject without providing a mechanism for rotating the camera.

In VAP, a prism having a variable apex angle is constructed by encapsulating, for example, silicon-based oil between two glasses connected by bellows and changing the relative angle between the two glasses. ing. The angle of refraction of the light flux passing through the VAP can be controlled by changing the angle between the two glasses.

Further, in the present proposal, since the optical axes of the two cameras can be rotated in the horizontal plane, one of the two glasses is fixed, and one glass is used as the center of the optical axis. It is rotatably supported about the direction of the rotation axis, and a voice coil motor and an angle detector are installed at the edge of the glass in the horizontal direction passing through the center of the optical axis. Is supplied. This VAP is provided on the most object side of the image pickup optical system. By thus eliminating the camera rotation mechanism, the two cameras can be installed side by side without a gap, and the device can be further downsized.

[0071]

As described above, according to the present invention, the first video camera section and the second video camera section in which the mechanism for setting the imaging conditions of each aperture, focus, and zoom are controlled by at least one control means. With the video camera section,
Uniting means for uniting the first and second video cameras,
Optical axis direction changing means for changing the optical axis directions of the first and second video cameras in a substantially horizontal plane with respect to a plane obtained by the optical axes of the first and second video cameras; A viewfinder displaying an image captured by at least one of the first and second video cameras, and a subject selecting unit provided in the viewfinder and capable of selecting an arbitrary subject in the viewfinder screen. In the compound-eye image pickup apparatus having the above-mentioned, since the optical axis angle of at least one of the first and second video cameras is controlled in accordance with the output of the subject selecting means, the stereoscopic shooting mode Not only the time, but also in the panoramic shooting mode, the observer can easily specify and select the subject while shooting.

[Brief description of drawings]

FIG. 1 is an external perspective view of a compound eye imaging device according to the present invention.

FIG. 2 is a block diagram showing the configuration of a compound eye image pickup apparatus according to the present invention.

FIG. 3 is a principle diagram of line-of-sight detection.

FIG. 4 is a flow chart showing initialization of the compound eye imaging apparatus according to the present invention when the apparatus is powered on.

FIG. 5 is a schematic diagram for setting an initial camera angle in a stereoscopic shooting mode.

FIG. 6 is a schematic diagram for explaining two shooting modes that can be realized when the visual axis detection is adopted in the compound eye imaging apparatus according to the present invention.

FIG. 7 is a diagram illustrating a cutout position of an image after a correlation process.

FIG. 8 is a diagram illustrating selection of an assect ratio in a panoramic shooting mode.

[Explanation of symbols]

 A first video camera section B second video camera section C combination means, 1,4 lens group 2,5 image sensor 15 control section 16 EEPROM 17, 18 convergence angle motor 19, 20 angle detector 24 viewfinder 26 line of sight Detector 33, 34 Distance detector 35 Stereoscopic / panorama switching SW 36 Synchronization signal generator 41, 44 Image memory 46 Image correlation processing unit 47 Image combining processing unit

Claims (1)

[Claims] 1. A first video camera unit and a second video camera unit in which a mechanism for setting an imaging condition of each aperture, focus, and zoom is controlled by at least one camera control unit, and the first video. A unit for joining the camera unit and the second video camera unit, and in a plane that is substantially horizontal to a plane required by each optical axis of the first video camera unit and the second video camera unit, Optical axis direction changing means for changing the optical axis directions of the first and second video camera sections, and optical axis angle control means for controlling the optical axis angles of the first video camera section and the second video camera section. A viewfinder for displaying an image captured by at least one video camera unit of the first video camera unit and the second video camera unit; and a viewfinder provided in the viewfinder, In a compound-eye image pickup device having a subject selecting unit capable of selecting an arbitrary subject in a viewfinder screen, in the first video camera unit and the second video camera unit corresponding to an output of the subject selecting unit. A compound-eye image pickup device, wherein an optical axis angle of at least one video camera unit is controlled.