JPH03192473A - Method and device for generating picture - Google Patents

Abstract

PURPOSE:To enable the picture processing of a picture equipped with a complicated shape or character in a short time as well by distributing a data processing to an arithmetic processing unit to execute an optimum three- dimensional picture processing method corresponding to the kind of shape and character data. CONSTITUTION:The shape and character data are stored in a storage 11 and out of the sent data, the secondary curved surface data, shadow or the data requiring reflection or diffraction are selectively received and parallelly processed by an arithmetic processing unit group 12, An arithmetic processing unit group 13 selectively receives the shadow or the polygonal data not to require the reflection or diffraction and parallely processes them and an arithmetic processing unit group 14 selectively receives the data to require the data processing of dispersed lighting and parallely processes them. Arithmetic results are three-dimensionally synthesized by a synthesizer 15 and one three- dimensional picture is generated and two-dimensionally displayed on the screen of a picture display device 16. Thus, the three-dimensional picture with high picture quality can be generated at high speed regardless of the picture to be generated.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image generation method and apparatus for computer graphics, etc., and particularly to an image generation method that can generate images of any shape at high speed, and Regarding equipment.

[Conventional collision]

When an image that represents a three-dimensional image on a two-dimensional surface is generated using a computer, etc., the processing amount of the processing unit becomes enormous and the calculation time becomes long, making it impractical.
For example, JP-A-62-84384 and JP-A-63-279.
As described in Publication No. 79, by providing a large number of processing units in parallel, allocating the problem of image generation processing to each processing unit and processing them in parallel, and finally combining the calculation results of each processing unit, Generating one image.

[Problem to be solved by the invention]

When trying to obtain arithmetic results in a short time, it is now common to allocate one problem to a large number of arithmetic processing units and process them in parallel. In image generation as well, one image can be processed in a short time by performing parallel processing.

However, as the shape of the object image to be generated and its properties such as the degree of transparency and light reflection of the object become more complex, the computational processing time becomes longer and less practical, no matter how much parallel processing is performed. .

An object of the present invention is to provide an image generation method and device that can process images of objects with complex shapes and properties in a short time, and a method for storing shape and property data suitable for processing with this device. There is a particular thing.

[Means to solve the problem]

The above purpose is to prepare a plurality of different processing methods as the processing method in image generation in which shape/property data is processed by an arithmetic processing device to create image data that represents a three-dimensional image on a two-dimensional surface. , distinguish the shape/property data to be processed for each processing method, distribute the discriminated shape/property data to the arithmetic processing unit that executes the processing method, and output each output of the arithmetic processing unit provided for each processing method. This is achieved by compositing and creating image data.

The above purpose is to store each shape/property data in a storage device in a shape/property data storage method.
This is achieved by adding information for discriminating the shape/property data or information indicating the optimal processing method for processing the shape/property data to each property data and storing the data in a storage device. .

Furthermore, the above object is achieved by adopting a configuration in which when data to be processed is read from a storage device and is passed to each arithmetic processing unit, the data is passed via a data buffer.

Further, the above object is achieved by performing processing only within polygons circumscribing the image when data processing is performed to generate an image.

[For production]

As three-dimensional image processing techniques, there are different techniques such as, for example, a two-buffer algorithm, ray tracing, and radiosity. Each of these methods differs in the shape and properties of the processing target that it specializes in. For example, the Z-buffer algorithm is fast in processing polygons, but processing them with functions slows it down, and furthermore, processing for adding reflections and shadows is impossible. Ray tracing can simulate light and process images to make it visible. Furthermore, radiosity can add a dull shadow to an image. However, there is a problem that the processing time is enormous.

In the present invention, one image processing problem is divided into multiple parts for each processing method. For example, an image generation part of only a simple shape is Z
Processing is passed to an arithmetic processing unit that executes a buffer algorithm, the image generation part that performs processes such as light reflection is passed to another arithmetic processing unit, and the image generation part that creates blurry shadows is further processed by another arithmetic processing unit. Since the processing is passed to the device and processed in parallel by each arithmetic processing device, it is possible to improve the processing speed, and it is also possible to improve the quality of the generated image. It goes without saying that even higher processing speeds can be achieved by performing the arithmetic processing that involves the image generation part, which takes one hour, in parallel using a large number of arithmetic processing units as in the past.

When storing shape/property data for image generation in memory, information and optimal processing methods are indicated according to the shape/property of the data type so that it can be allocated immediately when dividing image processing problems. By adding information, it is possible to read the data and immediately pass the data to an arithmetic processing unit that executes the optimal processing method, thereby improving processing speed.

In addition, by using data/<sofa, there is no disruption in data transfer and data transfer stagnation is avoided.
Even higher speed processing becomes possible.

Furthermore, by processing only the inside of the circumscribed figure, there is no wasted processing time for areas unrelated to one image generation.

〔Example〕

Hereinafter, one embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a configuration diagram of an image generation device according to an embodiment of the present invention. The image generation apparatus of this embodiment includes a storage device 11 storing shape/property data of a processing target, a processing unit group 12 for three-dimensional image generation processing method A, and an arithmetic processing unit for three-dimensional image generation processing method B. group 13, . . ., arithmetic processing unit group 14 for three-dimensional image generation processing method and an image display device 16 that displays a three-dimensional image on a two-dimensional screen.

As described above, there are multiple methods for projecting three-dimensional shape data onto a two-dimensional surface, but there is no single processing method that can process all shape data, property data, etc. at high speed and with high image quality. Therefore, if you perform image generation processing using a certain method,
The processing of shape data, etc., which this method is good at, can be processed at high speed and high image quality, but the processing of shape data, etc., which it is weak in, is delayed, resulting in unbalanced processing performance. However, in this embodiment, a plurality of processing methods are prepared. Processing method A
can process figures that can be expressed by quadratic surfaces, polygons, and other functions, and can also express shadows, reflections, refractions, etc. with high image quality. Processing method B can process only polygons.

Therefore, it is necessary to convert the quadratic surface into a polygon, which poses problems such as the speed of the conversion process and an increase in the amount of data for the conversion process. Furthermore, it is impossible to express reflection, refraction, etc. However, polygon graphic processing has the advantage that it can be processed much faster than processing method A. Processing method X is processing method A
In addition to the image display device, it has realistic expression capabilities such as diffused lighting. However, the processing speed is much slower than processing method A.

The storage device 11 for shape data and the like stores shape data such as quadratic curved surfaces and polygons, and property data such as the presence or absence of shadows, refraction and reflection, and whether diffuse illumination is to be performed. These data are sent to the arithmetic processing unit groups 12, 13, and 14, and the arithmetic processing unit group 12 selects quadratic surface data and data that require shadows, reflection, and refraction from the sent data. Take the target and process it in parallel. The arithmetic processing unit group 13 is
Polygonal data that does not require shadows, reflection, or refraction is selectively received and processed in parallel. The arithmetic processing unit group 14 selectively receives data that requires data processing of diffuse illumination, and processes the data in parallel. The calculation results from each processing unit group 12, 13 and 14 are three-dimensionally synthesized in a synthesis bag [15].
Three-dimensional images are generated and displayed two-dimensionally on the screen of the image display device 16.

According to this embodiment, compared to the case where all processing is performed using only processing method A, it is more efficient to process parts that can be processed using processing method B in parallel in the arithmetic processing unit group 13 and processing using processing method C. Since the parts that affect the image quality are processed by the arithmetic processing unit group 14, the overall processing speed is improved and even higher image quality can be achieved. In addition, compared to when all processing is performed using processing method B alone, it is possible to express shadows, reflections, refraction, etc., so it is possible to obtain high-quality images, while parallel processing is performed, so the processing speed is also lower. The effect is that you don't have to sacrifice that much. Furthermore, compared to the case where all processing is performed using processing method X, since parts that do not need to be performed using processing method X are processed by the arithmetic processing unit group 12, 13, faster processing is possible. Therefore, in this embodiment, images having complex shapes and properties can be obtained at high speed and with high quality.

In the above-described embodiment, the arithmetic processing unit group judges the data sent from the storage device 11 and decides whether or not it is data to be processed by itself, but in order to shorten this judgment time, the filter means This can be done by using For example, as shown in FIG. 2, data from the storage device 11 is passed through a shape data etc. identification filter (for processing method A) 21, a shape data etc. identification filter (for processing method B) 22, and a shape data etc. identification filter (for processing method B) 21. By passing the data through a filter (for processing method

In order to enable selection by such filtering means, information that can identify the data, such as a code indicating the type of shape or property, or an optimal processing method for processing, is added to the various data stored in the storage device 11. Attach information that indicates. As a result, the filter means 21, 22, 23 can easily sort the data based on the codes and information.

FIG. 3 is a block diagram of an image generation device using the above-mentioned filter means. Each arithmetic processing unit group 12, 13, 14 of this image generation device is provided with the above-mentioned filter 21, 22, 23 at its input stage, and the selected data is manually transmitted to a plurality of arithmetic processing units 24 for image data generation processing. , the calculation results of each arithmetic processing device 24 are collected by a control and data collection device 25, and the output thereof is output to a three-dimensional image synthesis device 15. This embodiment has the advantage that data selection processing can be performed reliably and at high speed.

The arithmetic processing unit group 43 of the image generation device according to the embodiment shown in FIG.
4 is provided with a buffer 42 such as a FIFO memory. Two-dimensional image generation processing methods include methods that sequentially process input shape data, etc.

There is a method Y in which processing is performed after all shape data etc. are collected. On the side that performs sequential processing, when the speed at which one piece of shape data or the like is processed becomes slower than the input speed of the shape data or the like, a stagnation occurs in the transfer of the shape data or the like. For this reason, the arithmetic processing unit 24 that executes the processing method @X and the processing method Y
If the arithmetic processing units 24 that execute Y are configured to operate in parallel, there is a problem in that the processing start of the arithmetic processing units that execute the processing method Y is delayed, resulting in an increase in processing time. However, by providing a buffer 42 and temporarily storing data from the filter means 41 as in this embodiment, it is possible to avoid data transfer stagnation.
Processing method X and processing method Y described above can be processed in parallel, and it is possible to improve the processing speed of the - layer.

The following methods can be considered to further improve the calculation processing speed. Normally, in an image generation device, calculations are performed on all pixels constituting both sides of one display. For this reason, calculations are also performed on the pixels in the area where there is no image to be displayed. Since this is a wasteful calculation, it is possible to improve the calculation processing speed by eliminating this waste. For example, as shown in FIG.
When displaying 1, calculations are performed only on pixels inside a polygon (rectangle in the illustrated example) 52 circumscribing the image 51. In the three-dimensional image generation processing method, -
Since the calculation processing time increases according to the number of pixels to be processed within the film, the processing time is shorter if the calculation is performed only on the pixels inside the rectangle 52 (shaded area) rather than on all the pixels on the screen 53. When each arithmetic processing unit is assigned data to be processed, it first calculates a polygon circumscribing the image as preprocessing, and performs arithmetic operations only on pixels within the polygon. This eliminates wasted processing time and enables even higher speed processing.

Note that the pixel range to be calculated by each processing unit is not determined by the main processing unit, which allocates calculation processing problems to each processing unit, but the main processing unit determines the calculation range and sends this to each processing unit for calculation along with the data. It may also be configured to specify a range.

In the embodiments described above, the types of arithmetic processing units were not described, but for example, all the arithmetic processing units 24 in FIG. Processing unit group 1 uses a dedicated processor designed to exclusively process method A.
The arithmetic processing unit 24 of the arithmetic processing unit 3 uses a dedicated processor designed to exclusively process processing method B, and the arithmetic processing unit 24 of the arithmetic processing unit group 14 is designed to exclusively process processing method X only. A configuration using a dedicated processor may also be used. Using all general-purpose processors has the advantage that it is possible to freely change which processing method and how many processors are used depending on the image to be generated. Further, by providing a dedicated processor, even faster processing becomes possible.

〔Effect of the invention〕

According to the present invention, it is possible to generate high-quality three-dimensional images at high speed regardless of the image to be generated.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an image generation device according to an embodiment of the present invention, and FIG. 2 is an explanatory diagram of a data selection method. FIG. 3 is a block diagram of an image generation device according to another embodiment of the present invention, FIG. 4 is a block diagram of a group of arithmetic processing units constituting an image generation device according to yet another embodiment of the present invention, and FIG. FIG. 2 is a diagram illustrating a range in which image processing is performed. DESCRIPTION OF SYMBOLS 11... Data storage device, 12... Arithmetic processing unit group for processing method A, 13... Arithmetic processing unit group for processing method B, 14... Arithmetic processing unit group for processing method X, 15...・
Three-dimensional image synthesis device, 16... image display device, 21.
22.23.41... Shape data etc. identification filter, 2
4... Arithmetic processing unit, 42... Buffer, 52...
- A polygon that circumscribes the image.

Claims (1)

[Claims] 1. An image generation method in which a plurality of arithmetic processing units are operated in parallel using a three-dimensional image processing method to create image data that expresses three-dimensional shape/property data on a two-dimensional surface. Different three-dimensional image processing methods are prepared, and data processing is assigned to an arithmetic processing unit that executes the three-dimensional image processing method most suitable for processing the shape/property data according to the type of the shape/property data. An image generation method characterized by generating an image by combining outputs of processing devices. 2. In an image generation method in which shape/property data is processed by an arithmetic processing device to create image data that represents a three-dimensional image on a two-dimensional surface, a plurality of different processing methods are prepared as the above-mentioned processing methods, and each processing method is The shape/property data to be processed is discriminated for each method, the discriminated shape/property data is distributed to the arithmetic processing unit that executes the processing method, and the outputs of the arithmetic processing units provided for each processing method are combined. An image generation method characterized by creating image data. 3. In an image generation method in which shape/property data is processed in parallel by multiple arithmetic processing units to create image data that represents a three-dimensional image on a two-dimensional surface, multiple different image generation processing methods are prepared, and each image generation process is A processing unit group is allocated for each method, and the shape/property data to be processed is divided into shape/property data suitable for each image generation processing method, and the processing is distributed to the processing unit groups to perform each image generation process. The method is executed in parallel by each processing unit group, and within each processing unit group, the shape/property data distributed to the processing unit group is processed in parallel, and the outputs of each processing unit are combined. An image generation method characterized by creating image data. 4. The image generation method according to any one of claims 1 to 3, wherein each arithmetic processing unit processes only within a polygonal region circumscribing the image data to be created. 5. Arithmetic processing unit a_1 that processes given shape/property data a using three-dimensional image generation processing method A, and arithmetic processing that processes given shape/property data b using three-dimensional image generation processing method B. The device b_1 and the shape/property data to be processed are divided into the shape/property data a and the shape/property data b.
A shape/property data discriminating means that separately supplies the shape/property data a to the arithmetic processing unit a_1 and the shape/property data to the arithmetic processing unit b_1, and synthesizes the respective outputs of the arithmetic processing unit a_1 and the arithmetic processing unit b_1. an image data synthesis device that generates three-dimensional image data and displays it on a display device. 6. Processing unit group a_1 that processes the given shape/property data a in parallel using the three-dimensional image generation processing method A, and processes the given shape/property data in parallel using the three-dimensional image generation processing method B. processing unit group b_1, and the shape/property data to be processed are combined with the shape/property data a.
Shape/property data a is sent to arithmetic processing unit group a_1 and shape/property data b is sent to arithmetic processing unit group b.
_1, and an image data synthesis device that synthesizes the respective outputs of the arithmetic processing unit group a_1 and the arithmetic processing unit group b_1 to generate three-dimensional image data and display it on a display device. An image generation device characterized by: 7. In a shape/property data storage method for storing shape/property data to be processed in a storage device when executing the image generation method according to any one of claims 1 to 4, each shape/property data is A method for storing shape/property data, characterized in that information for discriminating the shape/property data or information indicating an optimal processing method for processing the data is added and stored in a storage device. 8. A shape/property database storing shape/property data using the shape/property data storage method according to claim 7, a plurality of arithmetic processing devices each executing a different three-dimensional image generation method, and a three-dimensional image generation method. A filter means for discriminating the shape/property data to be assigned to each correspondingly provided arithmetic processing unit based on the information added to the shape/property data, and a filter means for synthesizing the outputs of each arithmetic processing unit to generate three-dimensional image data. An image generation device comprising: an image data composition device that creates an image data configuration device. 9. A shape/property database storing shape/property data using the shape/property data storage method according to claim 7, a plurality of arithmetic processing units each performing a different three-dimensional image generation method, and three-dimensional image generation. The shape/property data to be assigned to each processing unit group provided for each method is
An image generation device comprising: a filter means for making a discrimination based on the information added to the property data; and an image data composition device for creating three-dimensional image data by synthesizing the outputs of each arithmetic processing unit group. 10. The image generation device according to claim 5, claim 6, claim 8, or claim 9, wherein the arithmetic processing device is a general-purpose arithmetic processing device. 11. The image generation device according to claim 5, claim 6, claim 8, or claim 9, wherein the arithmetic processing device is a arithmetic processing device dedicated to a three-dimensional image generation processing method to be executed. 12. The image generating device according to claim 9, further comprising a shape/property data buffer provided between the filter means and the arithmetic processing unit group.