JPH07200870A - Stereoscopic three-dimensional image generator - Google Patents

Abstract

PURPOSE:To provide a stereoscopic three-dimesional image generator which can simplify its constitution and can reduce the entire computational complexity. CONSTITUTION:The processing parts prepared between a fluoroscopic projection conversion processing part 16 and a shading processing part 19 are used in common to the left and right images, the acquired raster data are shifted by a left-light shift processing part 20 for the left and right images based on the depth information on each pixel. These shifted raster data are shown on the left and right displays 23 and 24. At this time, a series of common processings are carried out with the visual points set at the centers of both eyes.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an improvement of a three-dimensional image generating apparatus for stereoscopic vision which presents different images to the left and right eyes based on parallax.

[0002]

2. Description of the Related Art A method of constructing a three-dimensional model of an object by numerical data in a computer and displaying the model three-dimensionally is important not only in the fields of numerical analysis and CAD but also in many application software. There is. Such a stereoscopic display has been performed by simply projecting a three-dimensional model in two dimensions and displaying it on a flat display, but in recent years, stereoscopic display has been achieved by presenting left and right visible images, respectively. The stereoscopic view to be displayed is realized.

There are the following two methods for stereoscopically presenting two images for such stereoscopic viewing. (1) A method for displaying an image for the left eye and an image for the right eye in a time-sharing manner on a single display, and showing them using glasses that open and close a liquid crystal shutter or the like in synchronization with the time-sharing. (2) A method in which two displays are prepared, and an image that can be seen from the left and right eyes is presented on each display so that they can be seen by each eye.

In any of the methods, in order to generate two stereoscopic images, conventionally, for example, as shown in FIG. 6, model data to be displayed is subjected to modeling conversion, and thereafter,
The perspective projection conversion, clipping processing, hidden surface removal and shading processing are performed independently for each of the left and right images. The obtained display data (raster data) is written in each corresponding frame buffer and then displayed individually on each display or displayed on one display in a time-division manner. In addition to this, there is also a configuration in which the left and right are all independent from the modeling conversion.

[0005]

As described above, in the conventional apparatus, a series of processes from perspective projection conversion to clipping processing, hidden surface removal processing, and shading processing are independently configured for the left and right displays. Is being processed. As shown in FIG. 7, when an object is viewed from above, the images viewed from the left and right eyes are displaced by the distance between the eyes, so the objects viewed by each eye This is because the shapes and colors of things look different.

However, in the above-described conventional three-dimensional image generation apparatus for stereoscopic vision, each image viewed from the left and right eyes is treated with an independent structure, so that it is twice as large as the case where one image is generated. However, there is a problem in that the device configuration is required. Further, when the above various processes are processed in one set on the left and right, for example, since the left and right images are sequentially generated and synchronized, the amount of calculation until the generation of one set on the left and right ends is large. There is a problem that the processing time becomes long.

The present invention has been made in view of the above problems of the prior art, and provides a stereoscopic three-dimensional image generation apparatus capable of simplifying the apparatus configuration and reducing the calculation amount as a whole. Is an issue.

[0008]

In order to solve the above problems, the present invention has the following constitution. That is, the present invention is
In a stereoscopic 3D image generation apparatus that generates an image for the left eye and an image for the right eye and displays the image for the left eye and the right eye, respectively, a 3D model data of a display object is placed at the center of the left and right eyes. Field coordinate conversion processing means for converting into a visual field coordinate system, stereoscopic effect realization processing means for adding processing for realizing a stereoscopic visual effect to the processing result of the visual field coordinate conversion processing means, and processing by the stereoscopic effect realization processing means A stereoscopic three-dimensional image generation device comprising: a left-right shift processing unit that adds left-right shift processing according to a difference in viewpoints to a result to generate an image for the left eye and an image for the right eye.

Further, in the present invention, the pixel data, which is the processing result of the stereoscopic effect realizing processing means, is shifted to the left or right in proportion to the depth direction from the viewpoint to generate an image for the left eye and an image for the right eye. A left-right shift processing means for generating can be provided. Further, in the present invention, the left-right shift processing means may include an interpolation means for performing interpolation between the shifted pixels.

[0010]

In the present invention configured as described above, one image recognized when the left and right images are viewed stereoscopically is
Utilizing the fact that it approximates the image obtained by placing the viewpoint at the center of the left and right eyes, the processing until the raster data is obtained is performed assuming that the viewpoint is placed at the center of the left and right eyes. That is, after the three-dimensional model data of the display object is converted by the visual field coordinate conversion processing means into a visual field coordinate system in which the viewpoint is placed in the center of the left and right eyes, the three-dimensional effect realization processing means processes the visual field coordinate conversion processing means. A process for realizing a three-dimensional visual effect is added to the result.

The left / right shift processing means adds a difference in viewpoint to the processing result (raster data) of the stereoscopic effect realizing processing means, that is, a difference between a temporarily set viewpoint (center of left and right eyes) and viewpoints of the left and right eyes. Add the left and right shift processing according to
Generate an image for the left eye and an image for the right eye. The left and right image data thus obtained are written in, for example, a frame buffer and displayed on the left and right displays.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, one embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a functional block diagram showing the configuration of one embodiment of a three-dimensional image generation apparatus for stereoscopic vision according to the present invention, FIG. 2 is a flowchart showing a stereoscopic image generation process in the same embodiment, and FIG. FIGS. 4 and 5 are diagrams for explaining the data shift, and FIGS. 4 and 5 are diagrams for explaining the left and right shift processing.

As shown in FIG. 1, the stereoscopic three-dimensional image generation apparatus of the present embodiment is roughly classified into a control unit 10 for comprehensively controlling each unit and transmitting three-dimensional model data of a display object. A geometry engine unit 11 that calculates coordinates, a raster engine unit 12 that generates raster data for left and right display, a display unit 13 for the left eye, and a display 14 for the right eye.

The three-dimensional model data sent from the control unit 10 is data that numerically represents a three-dimensional display object. The geometry engine unit 11 includes a modeling conversion processing unit 15 that performs modeling conversion, a perspective projection conversion processing unit 16 that performs perspective projection conversion, and a clipping processing unit 17 that performs clipping processing. The modeling conversion is a process of converting the coordinate system of the three-dimensional model data into the object coordinate system (absolute coordinate system). Further, the perspective projection conversion is a process of fixing the viewpoint and projecting an object based on the position and direction of the viewpoint, and is a process of converting the object coordinate system into the visual field coordinate system.

Conventionally, since the positions of the viewpoints of the left and right eyes are different, two perspective projection conversions for the left eye and the right eye have been performed as described above. One perspective projection conversion is performed with the position as the viewpoint. This is based on the fact that one image recognized when the left and right images are viewed stereoscopically approximates the image obtained by placing the viewpoints at the centers of the left and right eyes. Also,
The clipping process is a process of designating a display area on the display and removing the data that is padded outside.

The raster engine unit 12 also performs a hidden surface removal processing unit 18 for performing a hidden surface removal process, a shading processing unit 19 for performing a shading process, and a process for shifting the obtained raster data for left and right display. The hidden surface removal processing unit 18 is connected to a Z buffer 18a used for the hidden surface removal processing. The hidden surface erasing process is a process of erasing a surface (hidden surface) that is actually hidden and invisible in order to obtain a highly realistic image that is close to the real thing. The surface of the part to be displayed is removed, and only the part visible from the front is left and displayed.

In this embodiment, the so-called Z buffer method is used as a hidden surface erasing method. This is provided with a memory (the Z buffer 18a) that stores the distance in the Z-axis direction (depth direction) for each pixel (pixel), and when writing new pixel data, compares the old and new distance data, This is a method in which the pixel data closer to is left. In addition, the shading processing calculates how much the light hitting the model is reflected in a specific direction when displaying a model of a three-dimensional object, and shades the surface according to the direction and direction of the surface. It is a process to attach. When this shading process is performed, the interrelationships of objects are further clarified,
Reality increases. A stereoscopic visual effect is realized by such hidden surface removal processing and shading processing, and the image data (raster data) obtained by these processings have a stereoscopic effect.

The contents of the left and right shifting processing will be described in detail later, but the object is determined according to the difference between the tentatively set viewpoint (center of the left and right eyes in this embodiment) and the viewpoints of the left and right eyes. The image data for the left eye and the image for the right eye are generated in accordance with the depth value (Z value) of the raster data.

The left and right display units 13 and 14 are respectively composed of frame buffers 21 and 22 for storing data obtained by the left and right shift processing and displays 23 and 24 for displaying images. Frame buffer 2
Reference numerals 1 and 22 denote large-capacity memories for storing digital format image data, which read contents at a display speed,
Pixel data can be written simultaneously with / A conversion and display.

The coordinate conversion processing means is a perspective projection conversion processing section 16, and the stereoscopic effect realization processing means is a hidden surface removal processing section 18.
The shading processing unit 19 and the left / right shift processing unit are configured by a left / right shift processing unit 20, respectively.

Next, an image generation process by the present apparatus having the above-mentioned structure will be described with reference to the flow chart shown in FIG. First, when the control unit 10 sends the three-dimensional model data 30 (see FIG. 3A) of the display object to the modeling conversion processing unit 15 (S1), the modeling conversion processing unit 1
5 performs modeling conversion such as enlargement, reduction, rotation, and parallel movement on the model data 30 to calculate model data 31 (see FIG. 3B) in the object coordinate system, and performs perspective projection conversion of this data. It is sent to the processing unit 16 (S2).

Next, the perspective projection conversion processing unit 16 performs perspective projection conversion on the model data 31 based on the preset position and direction of the viewpoint, and the model data 32 in the visual field coordinate system (see FIG. c))),
This data is sent to the clipping processing unit 17 (S3).
At this time, as described above, conventionally, the perspective positions of the left and right eyes are different, so that two perspective projection conversions for the left eye and the right eye are separately performed, but in the present embodiment, the left and right eyes are converted. One perspective projection conversion is performed with the center position as the viewpoint. Therefore, the amount of data processing for perspective projection conversion, which occupies most of the data processing time, is reduced to about half.

Next, the clipping processing section 17 removes the data outside the display area designated in advance and adjusts the model data 32 to the display size, and then sends this data to the hidden surface removal processing section 18 (S4). Next, the hidden surface removal processing unit 18 performs hidden surface removal processing by the Z-buffer method on the model data 32 subjected to the clipping processing, and the data 33 in which the hidden surface that is actually hidden is deleted.
(See FIG. 3D) is calculated, and this data is sent to the shading processing unit 19 (S5).

Next, the shading processing section 19 performs shading processing on the data 33 which has been subjected to the hidden surface removal processing, and the data 34 in which the surface is shaded according to the direction and direction of the surface (see FIG. e)) is calculated, and this data is sent to the left / right shift processing unit 20 as raster data (S6). The model data 32 is sequentially given a stereoscopic effect by undergoing the hidden surface removal processing (S5) and the shading processing (S6), and finally changes to image data (raster data) 34 having a strong sense of reality and being close to a real thing. Will be done.

The raster data 34 is moved to the left and right by the processing unit 20.
When input to the left eye, the left and right shift processing unit 20 performs left and right shift processing on the raster data 34 in parallel for the left eye and the right eye, and the left eye image 35 (see FIG. 3F) and the right eye are processed. The respective images 36 (see FIG. 3 (g)) for use are generated (S7). That is, each pixel (pixel) data of the raster data 34 is converted into a depth value (Z value) according to the difference between the tentatively set viewpoint (center of the left and right eyes in this embodiment) and the viewpoints of the left and right eyes. , That is, Z in the viewpoint coordinate system
Shift in proportion to the axial value. As a result, the larger the Z value, the larger the shift amount.

The actual shift amount is an amount proportional to the distance between the viewpoint and the object, the size of the object, and the display size. For example, in the case of generating an image for the left eye, as shown in FIG. 4, when the ridge line A and the ridge line B are compared, the ridge line B has a larger Z value. The amount of shift to the left indicated by the middle arrow increases. When this is actually shifted and displayed, the ridge line B appears to be shifted to the left as shown in FIGS. 5 (a) and 5 (b).

The image 35 for the left eye generated in step 7
And the image 36 for the right eye are stored in the corresponding frame buffers 21 and 22 (S8, S9), and displayed on the left and right displays 23 and 24 (S10, S11). When this is viewed using a stereoscope or glasses for stereoscopic vision, the object can be stereoscopically viewed.

Therefore, according to the present embodiment, the perspective projection conversion processing unit 16, the clipping processing unit 17, the hidden surface removal processing unit 18, and the shading processing unit 19 are shared for the left and right images. Although the calculation amount is roughly halved and the left / right shift processing unit 20 is added, the device configuration is simplified as a whole, the calculation amount of the entire device can be reduced accordingly, and the processing time can be shortened.

In this embodiment, the hidden surface removal processing and the shading processing are used as the method of producing the stereoscopic effect. However, the shading processing is not limited to this, and other rendering processing such as ray -Tracing (ray tracing method) may be used. In this embodiment, the left-eye image and the right-eye image are displayed on separate left and right displays, but one display may be used in a time-sharing manner, and different images may be displayed on the left and right eyes. Obviously, any presentation means that can be presented may be used.

[0030]

As described above, according to the present invention, a series of processes from the coordinate conversion process to the three-dimensional model data of the display object to the stereoscopic effect realization process are shared for the left and right images. The device configuration is simplified, the calculation amount of the entire device is reduced, and the processing time is shortened.

[Brief description of drawings]

FIG. 1 is a functional block diagram of an embodiment of a stereoscopic 3D image generation apparatus of the present invention.

FIG. 2 is a flowchart showing a stereoscopic image generation process in the embodiment.

FIG. 3 is an explanatory diagram showing a change in data after various kinds of processing using a cube as an example.

FIG. 4 is an explanatory diagram of a shift amount for obtaining an image for the left eye in the left and right shift processing.

5 is a diagram showing an image shifted according to the shift amount of FIG.

FIG. 6 is a functional block diagram showing a data flow of a conventional three-dimensional image generation apparatus for stereoscopic vision.

FIG. 7 is an explanatory diagram showing a state where an object is viewed from the left and right eyes.

[Explanation of symbols]

 10 Control Unit 15 Modeling Conversion Processing Unit 16 Perspective Projection Conversion Processing Unit (Coordinate Conversion Processing Unit) 17 Clipping Processing Unit 18 Hidden Surface Elimination Processing Unit (Stereoscopic Effect Realization Processing Unit) 19 Shading Processing Unit (Stereoscopic Effect Realization Processing Unit) 20 Left and Right Shift processing unit (left and right shift processing means) 21, 22 Frame buffer 23, 24 Display

Claims (3)

[Claims] 1. A stereoscopic 3D image generation apparatus for generating an image for the left eye and an image for the right eye and displaying the image for the left eye and the image for the right eye, respectively. A visual field coordinate conversion processing unit for converting into a visual field coordinate system with a viewpoint in the center, a stereoscopic effect realization processing unit for adding a process for realizing a stereoscopic visual effect to the processing result of the visual field coordinate conversion processing unit, and the stereoscopic effect. A three-dimensional image for stereoscopic vision, comprising: left-right shift processing means for adding left-right shift processing according to a difference in viewpoint to the processing result of the realization processing means to generate a left-eye image and a right-eye image. Generator. 2. The image data for left eye and the image for right eye according to claim 1, wherein the pixel data which is the processing result of the stereoscopic effect realization processing means is shifted to the left or right in proportion to the depth direction from the viewpoint. A three-dimensional image generation apparatus for stereoscopic vision, comprising a left-right shift processing means for generating 3. The stereoscopic three-dimensional image generation device according to claim 1, wherein the left-right shift processing unit includes an interpolation unit that interpolates between the shifted pixels.