JPH02267676A - Stereoscopic image device - Google Patents

Abstract

PURPOSE:To easily and accurately measure the spatial position of a body in a stereoscopic image by calculating the spatial position by retrieving a correspondence area centering the position of a cursor. CONSTITUTION:A three-dimensional graphic control circuit 13 outputs the three- dimensional position of the cursor K to a focus control circuit 14, and the focus control circuit 14 controls two video cameras 2 and 3 so as to obtain the focusing of the cameras 2 and 3 on a supplied three-dimensional position. Furthermore, an instruction to calculate the spatial position is issued to an image processing processor 15 via an instruction means such as a keyboard, etc. The image processing processor 15 stores the photographing images of the video cameras 2 and 3 in image memories 16 and 17, respectively, and sets a reference window W0 in the periphery of the cursor Kf at a left image 2a, and the correspondence area (e) centering the position of the cursor Kr at a right image 3a. In such a way, since the desired part of a targeted body exists in the correspondence area (e), it is enough to retrieve only the correspondence area, and the measurement of the position of the body easily and accurately is performed by using the stereoscopic image.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a stereoscopic imaging device, and more particularly to a stereoscopic imaging device that obtains a stereoscopic image using a video camera of a meeting.

[Conventional technology]

Conventionally, video cameras at a meeting were installed side by side, images from each video camera were switched in short cycles and displayed alternately on both sides of the monitor, and the liquid crystal shutter was driven in synchronization with the switching of images from the left and right video cameras. However, a stereoscopic image device is known that obtains a stereoscopic image by making the image of the left video camera visible only to the left eye and the image of the right video camera visible only to the right eye.

[Invention tries to solve! ! ! ! [Problem] The conventional stereoscopic imaging apparatus described above has a problem in that it can only observe a stereoscopic image and cannot measure the spatial position of an object within the stereoscopic image.

Therefore, it is an object of the present invention to provide a stereoscopic imaging device that can easily and accurately measure the spatial position of an object within a stereoscopic image.

(Means for Solving the Problems) The stereoscopic imaging device of the present invention switches the images of the video cameras of the meeting arranged side by side in a short cycle and displays them alternately on both sides of the monitor, and selectively displays the images of the left video camera by the selective viewing means. In a stereoscopic imaging device, the image of the right video camera is made visible only to the left eye, and the image of the right video camera is made visible only to the right eye. A left-view cursor image that is hypothetically obtained when a certain cursor is viewed with the left video camera is calculated, and a right-view cursor image that is hypothetically obtained when the same cursor is viewed with the right video camera. cursor image calculating means for calculating a spatial position; image combining means for combining the left-viewing cursor image with the image of the left video camera and combining the right-viewing cursor image with the image of the right video camera; a reference window for setting a reference window around the cursor image in the image of one of the left or right video cameras; window setting means; The predetermined mapping sets a region, the mapping sets a window within the region, and the mapping moves the window within the region to meet the criteria. The present invention is characterized in that it includes a matching means for performing matching with a window, and a spatial position calculating means for calculating a spatial position from the relative position of the reference window and the corresponding window when the two windows most closely match each other. It is something to do.

[Effect]

In the stereoscopic image device of the present invention, when the three-dimensional position of the cursor is designated, the cursor is displayed at a position in the stereoscopic image that corresponds to the designated three-dimensional position.

Then, select the image part that most closely matches the image part of the reference window set around the cursor in one image.
The correspondence approximately centered on the cursor in the other image is detected within the area, and the spatial position is calculated.

Therefore, by placing the cursor on an object in a stereoscopic image, the spatial position of the object can be obtained.

Here, since there is a limit to human resolution ability, even if the object and the cursor appear to match, they may actually be out of alignment. Since the spatial position is calculated by searching for , the spatial position obtained even if it is off, the positive value is 1!
Become something.

〔Example〕

Hereinafter, the present invention will be explained in more detail based on embodiments shown in the drawings. Here, FIG. 1 is a block diagram of a stereoscopic image device according to an embodiment of the present invention, FIG. 2 is a conceptual diagram showing the relationship between the three-dimensional position of the cursor and the cursor image, and FIG. 3 is a reference window and a window corresponding to the cursor image. FIG. 4 is a conceptual diagram for explaining the correspondence area, and FIG. 5 is a flowchart showing the order of spatial position ¥1 calculation. Note that the present invention is not limited to the embodiments shown in the figures.

In the stereoscopic imaging device 1 shown in FIG. 1, meeting video cameras 2.3 are installed side by side with their optical axes parallel to each other.

The image signal output from the video camera 2.3 is input to the field switching circuit 4.

The field switching circuit 4 switches the images from both cameras 2.3 into one based on the synchronization signal from the camera controller 5.
The signals are alternately switched for each field and output to the superimpose circuit 6.

The superimpose circuit 6 combines a left-view cursor image provided from a three-dimensional graphic control circuit 13 (described later) with an image of the left video camera 2, and also combines a right-view cursor image with an image of the right video camera 3. and outputs it to the double-speed scan converter 7.

The double-speed scan converter 7 converts the scan rate of the output image of the superimpose circuit 6 into a double scan rate and sends it to the display 8 in order to prevent flickering of the image.

The display 8 is a monitor TV capable of displaying images at double speed, that is, at a field period of 1/120 seconds.

The field switching circuit 4 switches between the image of the left video camera 2 and the image of the right video camera 3 for each field, and outputs a switching timing signal to the liquid crystal shutter control circuit 9A. Therefore, the liquid crystal shutter control circuit 9 switches the liquid crystal shutter 10 every time the image of the left video camera 2 and the image of the right video camera 3 are switched, and when displaying the image of the left video camera 2, only the left circularly polarized light is transmitted. However, when displaying an image from the right video camera 3, only right circularly polarized light is transmitted.

The observer uses polarized glasses 11;
The left eye of No. 1 has a lens that transmits only left-handed circularly polarized light, and the right eye has a lens that transmits only right-handed circularly polarized light. Only the image from camera 3 will be visually recognized. Therefore,
This will allow you to obtain a three-dimensional image.

The three-dimensional joystick 12 is used to specify the three-dimensional position (X, Yk, Z,) of the cursor, and the tilt and rotation state of the lever determines the three-dimensional position (Xm, Yh,
, Zm).

The three-dimensional graphic control circuit 13 reads the state of the three-dimensional joystick 12, obtains the three-dimensional position of the cursor (CXw, Yh, Za), and outputs a left-view cursor image and a right-view cursor image.

FIG. 2 shows the principle of this cursor image calculation. For convenience of explanation, the three-dimensional position of the cursor (Xh, Y
h, Zk) represents the point K, and the spatial coordinate value of the lens center Ql of the left video camera 2 is [Xo.

Y,,Z,), and the center of the lens of the right video camera 3 is Q1.
The spatial coordinates of are (X0+L, Y, , Z,), and the distance between both video cameras 2.3 is f.

Also, images 2. of both video cameras 2 and 3 are shown. ．． 3. The upper coordinate origin is each image 2. ．． 3. and each optical axis.

Then, according to the following equation, the image 2 of the left video camera 2 at the point
．． The coordinates (χ1
, y) are output. Similarly, the image 3 of the right video camera 3 at the point. 1 image of 2 at the corresponding point above 3. The related coordinates (xr, y) are calculated.

xs ==(Xh Xo) r/ (Za
Za )x--(X-10L Xh) f/ (Z
+= Z-)y-(Yi= Y-)r/ (Zk
Zo) Above image 2. ．． 3. At the upper point, the image of the left video camera 2 and the right video camera 3 are shown with Kr as the shining point.
When combined with the image of
A point K that shines at a spatial position in the stereoscopic image corresponding to Yk, Zk) is recognized. Therefore, by expanding this, a three-dimensional cursor K having a desired shape can be displayed in a three-dimensional image.

This cursor can be freely moved within the stereoscopic image by operating the three-dimensional joystick 12.

Returning to FIG. 1, the three-dimensional graphic control circuit 13 controls the cursor's three-dimensional position 1 (xt,).

Yk, Zk) are output to the focus control circuit 14.

A focus control circuit 14 controls both cameras 2 and 3 so that the video camera 2.3 focuses on a given three-dimensional position (Xb, Y, zk).

Therefore, if the observer operates the three-dimensional joystick 12 and moves the cursor K to the desired part of the stereoscopic image,
The video camera 2.3 will focus on that point. In this way, it becomes possible to easily adjust the focus of the video camera 2.3 while observing the stereoscopic image.

Now, the three-dimensional position (Xb, Y*) of the cursor.

Since zm) is known, by aligning the cursor K with the desired part of the target object, the three-dimensional position of the desired part can be found. However, it is difficult for humans to accurately align the cursor K with a desired portion of the target object, and a certain amount of deviation is inevitable. Therefore, although the approximate position of the desired portion of the target object is known as described above, the exact position is not known.

Therefore, if an instruction to calculate a spatial position is given to the image processing processor 15 via an instruction means such as a keyboard, for example,
The image processor 15 stores the images taken by the video camera 2.3 in the image memory 16.17, respectively.

Next, the image processing processor 15 sets a reference window W0 around the cursor + in the left image pane, as shown in FIG.

Also, right image 3. The correspondence area e is set for the cursor in , centered on the position of . This correspondence wI area e
As shown in FIG. 4, is an area on the image that corresponds to the error range E that occurs before and after the cursor. Therefore, since the desired portion of the target object is located within the area e, it is sufficient to search only within the area for this association.

As shown in FIG. 3, the image processing processor 15 sets the mapping window Wo at the left end of the area e, and
The tH1u characteristic with the reference window W is determined. In this process, for example, feature points of both windows W, , W, are extracted by differential processing or the like, and plate matching is performed.

Next, the mapping moves the window Wx by one pixel to the right along the Epipolar line and performs the same processing as above, and the mapping repeats this until it reaches the right end of the area e.

When the position with the highest similarity is detected in this way, the spatial position is calculated from the positions of both windows W, , W, and l at that time and displayed on the CRT.

In this way, it becomes possible to know the exact spatial position.

FIG. 5 shows the above processing in flow steps.

In the above explanation, the optical axes of the video camera 2.3 are assumed to be parallel, but the present invention can be similarly applied even when the angle of convergence is taken into account.

〔Effect of the invention〕

According to the present invention, images from video cameras of a meeting arranged side by side are switched at short intervals and alternately displayed on both sides of the monitor, and the image from the left video camera is viewed only with the left eye by the selective viewing means, and the images from the right video camera are displayed on both sides of the monitor alternately. In a stereoscopic imaging device in which a camera image is viewed only by the right eye L2, a cursor position specifying means for specifying the three-dimensional position of the cursor, and when the cursor at the specified three-dimensional position is viewed with the left video camera. cursor image calculation means for calculating a left-looking cursor image that is hypothetically expected to be obtained when the cursor is viewed with a right video camera; image combining means for combining the left-viewing cursor image with the image of the left video camera and combining the watching cursor image with the image of the right video camera; instruction means for instructing to calculate a spatial position; Alternatively, a reference window setting means for setting a reference window around the cursor image in the image of one video camera on the right, and a predetermined correspondence setting area approximately centered on the position of the cursor image in the image of the other video camera. Attached is an area setting means, a matching means for setting a corresponding window in the corresponding area, and moving the corresponding window within the corresponding area to match it with the reference window; A stereoscopic image device is provided, characterized in that the reference window and the correspondence include a spatial position calculation means for calculating a spatial position from the relative position of both windows when the windows most closely match each other. It becomes possible to easily and accurately measure the object position using the .

[Brief explanation of drawings]

Fig. 1 is a block diagram of a stereoscopic imaging device according to an embodiment of the present invention, Fig. 2 is a conceptual diagram showing the relationship between the three-dimensional position of the cursor and the cursor image, and Fig. 3 shows the reference window and the correspondence between the windows. FIG. 4 is a conceptual diagram for explaining the correspondence and 61 areas, and FIG. 5 is a flow diagram showing the order of spatial position calculation. [Explanation of symbols] 1...Stereoscopic image device 2...Left video camera 3...Right video camera 4...Field switching circuit 6...Superimpose circuit 8...Display 9...- 2 night shutter control circuit 10...
Liquid crystal shutter 11...Polarized glasses 12...Three-dimensional joystick 13...Three-dimensional graphic control circuit 14.
...Focus control circuit 15...Image processing processor 16.17...Image memory K...Cursor e...Associated with area Wo...Reference window WX...Associated with window.

Claims (1)

[Claims] 1. Images from video cameras of a meeting arranged side by side on the left and right are switched in short cycles and alternately displayed on both sides of the monitor, and a selective viewing means allows images from the left video camera to be viewed only with the left eye. In a stereoscopic imaging device in which the image of the right video camera is made visible only to the right eye, there is provided a cursor position specifying means for specifying the three-dimensional position of the cursor, and a means for specifying the three-dimensional position of the cursor, and a means for specifying the three-dimensional position of the cursor as viewed by the left video camera. cursor image calculation means for calculating a left-viewing cursor image that is hypothetically obtained when the cursor is viewed with a right-viewing video camera; image synthesis means for synthesizing the left-view cursor image with the image of the left video camera and the right-view cursor image with the image of the right video camera; instruction means for instructing to calculate a spatial position; Or, a reference window setting means for setting a reference window around the cursor image in the image of one video camera on the right, and a correspondence for setting a predetermined correspondence area approximately centered on the position of the cursor image in the image of the other video camera. area setting means; matching means for setting a mapping window within the mapping area and moving the mapping window within the mapping area to match the reference window; 1. A stereoscopic image device comprising: spatial position calculation means for calculating a spatial position from the relative position of the attached windows when they most closely match.