JP2003109033A - Three-dimensional graphics drawing device for performing occlusion culling - Google Patents

Abstract

PROBLEM TO BE SOLVED: To more accurately and fast perform occlusion culling in a three- dimensional computer graphics system. SOLUTION: The number of polygons of each object is reduced in accordance with the ratio of the each object to the size of an assembly model B0 consisting of objects B1 to B6, and the object in which the number of polygons is reduced is used to perform occlusion decision. Then, an object decided not to be occluded is displayed by using polygons that are more than polygons used in the occlusion decision.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a drawing device for drawing a three-dimensional graphics model constructed on a computer.

[0002]

2. Description of the Related Art Today, mechanical CAD (computer-aided d)
3D computer graphics (3DCG) is used in various fields such as virtual reality systems such as esign) systems, graphics games, animations, walkthroughs and the like.

One of the conventional methods for drawing 3DCG at high speed
As one, occlusion cullin
g) Law is known (USP 5,751,291 May / 1998 Olsen
et al., USP 5,613,050 Mar / 1997 Hochmuth et al.
). In this method, a CG model (object) of an object to be drawn is covered with a basic figure such as a rectangular parallelepiped, and the basic figure is used instead of the object to determine whether or not each basic figure is occluded by another basic figure. judge.

Here, being occluded by another figure means that the figure is hidden behind another figure. By omitting the drawing of the objects in the occluded basic figure, the drawing speed is increased.

[0005]

However, the above-mentioned conventional occlusion culling method has the following problems.

In this method, since the object is covered with a basic figure such as a rectangular parallelepiped, in the case of an object having a complicated shape, a situation in which one of the objects is a shadow of the other is overlooked,
On the contrary, it is considered that the drawing of an object that does not become a shadow is often omitted. Therefore, it can be said that this method is suitable for a system such as a walkthrough that handles an object having a relatively simple shape.

However, in a mechanical system design site, an object having a complicated shape consisting of several hundreds of thousands of polygons often appears, and the occlusion culling method based on the basic figure cannot be expected to draw accurately. Therefore, an occlusion culling method capable of drawing a 3D model composed of a large number of polygons at high speed, increasing the update rate of the computer screen, and smoothly rotating and moving the model is desired.

SUMMARY OF THE INVENTION An object of the present invention is to provide a drawing apparatus which performs occlusion culling more accurately and at high speed in a CG system that handles complicated shapes such as mechanical CAD.

[0009]

FIG. 1 is a principle diagram of a drawing apparatus of the present invention. According to the first principle of the present invention, the drawing apparatus includes the shape information storage unit 1, the table unit 2, the virtual image storage unit 3, the determination unit 4, and the occlusion culling unit 5, and displays a plurality of objects.

The shape information storage means 1 stores the shape information of a plurality of objects, and the table means 2 stores a plurality of color information virtually associated with the plurality of objects in a one-to-one correspondence. The virtual image storage means 3 stores virtual color image information drawn based on the information in the shape information storage means 1 and the table means 2. The determination unit 4 scans the virtual color image information to perform occlusion determination, and the occlusion culling unit 5 omits the display of the object determined to be occluded among the plurality of objects.

The table means 2 stores a one-to-one correspondence between object identification information and color information. The color information virtually associated with the object is not the color information for actually displaying the object, but the color information for virtually drawing the object for occlusion determination. Virtual color image information drawn using this virtual color information is not actually displayed on the screen.

The determination means 4 scans the virtual color image information stored in the virtual image storage means 3 in pixel units to extract color information, and refers to the table means 2 to correspond to the extracted color information. Recognize objects. Then, those objects are determined to be unoccluded objects, and the other objects are determined to be occluded objects. Occlusion Culling Means 5
Hides objects determined to be occluded.

The point of the first principle of the present invention is to generate image information by using color information uniquely assigned to each object, and perform occlusion determination based on the image information.

According to such a drawing apparatus, since the shape of the object itself is used instead of using the basic figure that covers the object, more strict occlusion determination can be performed. Further, by scanning the image information drawn using the color information uniquely assigned to each object, the determination can be performed at high speed.

Further, according to the second principle of the present invention, the drawing device is provided with the judging means 4, the reduction means 6, and the display means 7, and displays a plurality of objects constituting the three-dimensional graphics model.

The reduction means 6 reduces the number of geometry primitives forming each object according to the ratio of the size of each object to the size of the three-dimensional graphics model, and the determination means 4 determines Occlusion determination is performed using an object with a reduced number of geometry primitives. The display means 7 displays the objects determined not to be occluded among the plurality of objects by using more geometry primitives than the number of geometry primitives used for the occlusion determination.

The geometry primitive is geometric data representing a part of the surface of the object, and, for example, a triangular polygon corresponds to this. The reduction unit 6 generates a model in which the number of geometric primitives is reduced for each object. The determination unit 4 performs the occlusion determination using the model generated by the reduction unit 6, and the display unit 7 displays the object determined not to be occluded by using, for example, the geometry primitive before being reduced. .

The point of the second principle of the present invention is to reduce the number of geometric primitives of each object to perform occlusion determination. According to such a drawing apparatus, since the number of geometry primitives is reduced and the shape of the object is simplified, the occlusion determination can be performed at high speed. Further, since the shape of the model in which the number of geometric primitives is reduced reflects the shape of the original object, it is possible to make a strict determination.

For example, the shape information storage means 1, the table means 2, and the virtual image storage means 3 of FIG. 1 correspond to a RAM (random access memory) 27 of FIG. 3 described later,
The determination means 4, the occlusion culling means 5, and the reduction means 6 in FIG. 1 correspond to the CPU (central processing unit) 23 in FIG. 3, and the display means 7 in FIG. 1 corresponds to the graphics processing device 25 and graphics in FIG. Corresponds to the display 26.

[0020]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described in detail below with reference to the drawings. In this embodiment, in order to determine whether or not an object is occluded by another object, different virtual color information (virtual color) is given to each object, and the object is identified using the color. Draw on the bitmap. Then, by scanning the color on the bitmap, it is determined whether or not an object is occluded by another object.

According to such occlusion culling determination, since the shape of the object itself is used instead of using the basic figure covering the object, more rigorous determination can be performed. Further, by scanning the color uniquely assigned to each object on the bitmap, the determination can be performed at high speed.

Occlusion culling is made more efficient by drawing the virtual color information only once every several frames. Further, by using a model subjected to polygon reduction as a model for drawing virtual color information, occlusion culling is further speeded up.

FIG. 2 is a block diagram of a drawing apparatus for drawing an assembly model created by mechanical CAD or the like at high speed. The drawing apparatus of FIG. 2 includes a shape / assembly information storage unit 11, a position / orientation information update unit 12, an occlusion culling determination unit 13, a drawing unit 14, and a polygon reduction calculation unit 15.

The shape / assembly information storage unit 11 stores graphic information such as a set of vertices for expressing objects, a parent-child relationship expressing an assembly relationship between objects, a position, a posture, and a joint relationship. Position / posture information update unit 1
2 calculates the amount of change with time of the position of each vertex representing the position / orientation of the object when the object moves.

The occlusion culling determination unit 13 draws each object with virtual color information, extracts the result into a bitmap, and detects the color appearing on the bitmap. This will extract the objects (non-occluded objects) that appear in the table without shadowing other objects and mark those objects.

The drawing unit 14 draws only the object marked by the occlusion culling judging unit 13 for several cycles. Polygon reduction calculator 15
The number of polygons of the object is reduced stepwise, and the object with the reduced number of polygons is input to the position / orientation information updating unit 12. These objects are used to speed up occlusion culling decisions.

FIG. 3 shows an example in which the drawing apparatus of FIG. 2 is configured by using an information processing apparatus (computer). The information processing device of FIG. 3 includes a data input device 21, interfaces (I / F) 22, 24, and CPU (central processing unit) 2
3, graphics processing device 25, graphic display 26, RAM (random access memory) 27,
And a storage device 28.

The CPU 23 updates the position / orientation associated with the movement of the object (coordinate conversion), occlusion culling, polygon reduction, calculation for displaying the calculation result, calculation for graphically displaying the motion of the object, and other operations. Performs all logical operations. Since the process of updating the position / orientation of an object in space can be regarded as the process of converting the coordinate system of the object into another coordinate system, such a process is often called the coordinate conversion of the object.

The storage device 28 requires a working environment, object shape data, assembly data, initial positions thereof, an execution module (program) for drawing algorithms, an execution module for graphically displaying the movement of an object, and the like. Data and execution modules are stored. Examples of the storage device 28 include a magnetic disk device, an optical disk device, and a magneto-optical disk (ma).
gneto-optical disk) device or the like is used.

The RAM 27 includes a shape data / assembly data memory 31, a coordinate conversion data memory 32, a virtual color information memory 33, an occlusion culling memory 34, a polygon reduction memory 35, and an arrangement table memory 36.

Shape data / assembly data memory 31
The data of the work environment and the object called from the storage device 28 by the CPU 23 are stored in the C, and the coordinate conversion data memory 32 is stored in the C by using the movement command data of the object input through the data input device 21.
The shape data of the object calculated by the PU 23 is held.

The virtual color information memory 33 holds a drawing result as a bitmap when different virtual colors are assigned to the respective objects in order to perform occlusion culling. Occlusion Culling Memory 34
Holds a flag indicating that the object having the color drawn on the bitmap of the virtual color information memory 33 is in the "table" state not occluded by any other object.

A model in which the number of polygons of each object is reduced resides in the polygon reduction memory 35. An object in a CG system is usually formed by a set of geometry primitives such as triangular polygons.

The model in the polygon reduction memory 35 includes graphic information such as true color, vertex coordinates, and normal vector of each geometric primitive forming an object. The highest level data in polygon reduction is the geometry primitive data of the original object, and the low level data is generated stepwise from this original geometry primitive data.

The arrangement table memory 36 holds a one-to-one correspondence between the identification information of the object and the virtual color. By referring to the array table memory 36, it is possible to obtain the virtual color corresponding to each object and the identification information of the object corresponding to each virtual color.

The data input device 21 sends the work environment, the position / orientation of the object included in the assembly model, the movement command data, etc. to the CPU 23 via the I / F 22. The CPU 23 updates the position / posture of each object while accessing the RAM 27 and the storage device 28.
Performs processing such as polygon reduction and occlusion culling.

The graphics processing device 25 is, for example,
The graphics board includes a plurality of frame buffers that temporarily store image information and a swapping mechanism that swaps the frame buffers and sends the image information to the graphic display 26.

When the occlusion culling is performed, the image information of only the objects in the table state is displayed on the frame buffer based on the information of the occlusion culling memory 34 and the polygon reduction memory 35 sent via the I / F 24. Be expanded. Then, the image information is sent to the graphic display 26 by the swapping mechanism and displayed on the screen.

The shape / assembly information storage unit 11 of FIG. 2 corresponds to the storage device 28 of FIG. 3, and the drawing unit 14 of FIG. 2 corresponds to the graphics processing device 25 and the graphic display 26 of FIG. Further, the position / orientation information updating unit 12, the occlusion culling determination unit 13, and the polygon reduction calculation unit 15 in FIG. 2 correspond to the execution modules stored in the storage device 28 in FIG. 3.

FIG. 4 shows a computer-readable recording medium capable of supplying a program and data to the information processing apparatus of FIG. When the program and data necessary for processing are stored in the portable recording medium 41, the information is loaded into the RAM 27 and then the CPU 23
Used by. As the portable recording medium 41, a memory card, a floppy (registered trademark) disk, a CD-RO
M (compact disk read only memory), optical disk,
Any computer-readable recording medium such as a magneto-optical disk is used.

When the program and data required for processing are stored in the external database 42 connected by the network, those information are stored in the RAM via the line.
After being loaded into 27, it is used by the CPU 23.

FIG. 5 is a flow chart of the drawing process by the drawing apparatus of FIG. First, the CPU 23 stores in the shape data / assembly data memory 31 from the storage device 28,
The shape data, the assembly data, and the initial position of the object included in the assembly model to be displayed are read (step S1). Next, an arbitrary polygon reduction algorithm is applied to each read object to create a model in which the number of polygons is generally reduced to M stages, and the model is stored in the polygon reduction memory 35 (step S2).

As a polygon reduction algorithm, for example, a quadric error metric
ic) algorithm is used (Michael Garl
and and Paul S. Heckbert, “Surface Simplification
Using Quadric Error Metrics, ”Proceedings of SIG
GRAPH 1997, pp.209-216, 1997.).

FIG. 6 shows an example in which the shape model of an object (sofa) having an initial polygon number of about several thousand is reduced to four levels. Here, the drawing level (level of
The model whose level (l, LOD) is level 0 corresponds to the original object, and the number of polygons decreases as the level decreases to level 1, level 2 and level 3.

Next, the CPU 23 controls the data input device 21.
The position and orientation of the read object are changed by a slight amount in accordance with the movement command data input from, and the changed shape data is stored in the coordinate conversion data memory 32 (step S3). When the user specifies the movement of the object, the movement command data is input by interactively moving the object with an input device such as a mouse. Further, when the system automatically generates the movement of the object, the movement command data is generated by a predetermined algorithm.

Next, each object is drawn in a different virtual color on the frame buffer in the graphics processing device 25, and a bitmap corresponding to the display screen is generated (step S4). Next, the generated bitmap is read into the virtual color information memory 33, the bitmap is scanned, and the drawn color is extracted. Then, referring to the array table memory 36, the objects corresponding to the extracted colors are recognized, and the flags of those objects in the occlusion culling memory 34 are set.

FIG. 7 shows an example of such occlusion culling determination. In FIG. 7, different virtual colors VC1, VC2, VC3, and VC4 are assigned to four objects A, B, C, and D having LOD of level 1. It is assumed that the state when these objects are viewed from the viewpoint P in the direction of the arrow is drawn on the frame buffer.

At this time, the objects seen from the viewpoint P are only a part of the objects A and B (the left half of the sofa). Object C and Object D
Both are hidden by the shadows of other objects and do not appear in the framebuffer. Therefore, two virtual colors VC1 and VC2 and a background color are extracted from the generated bitmap.

Here, if the background color is always white and this color is not assigned to the virtual color of another object,
Objects and virtual colors always have a one-to-one correspondence.
Therefore, at this moment, it can be concluded that only two objects A and B are visible, and two objects C and D are occluded by other objects. Therefore, the CPU 23 sets flags of the object A and the object B, and the object C
And turn off the flag of object D.

Next, the CPU 23 extracts the flagged objects in the occlusion culling memory 34, and the graphics processing unit 25 draws them in the frame buffer in their original colors and displays them on the graphic display 26. Yes (step S5). Then, the drawing device displays a state in which the assembly model dynamically changes by repeating the loop process including steps S3 to S5.

The virtual color on the frame buffer used for the occlusion culling judgment is erased when the judgment is completed and is not displayed on the display. Only the objects determined to be in the table state are actually displayed on the display.

For example, the drawing state at the moment immediately after the occlusion culling judgment as shown in FIG. 7 is as shown in FIG. In FIG. 8, only the objects A and B in the state of the table are drawn at the level 0 corresponding to the initial polygon number.

Next, when each object moves by a very small amount, the drawing state changes as shown in FIG. In FIG. 9, object C and object D are still occluded and therefore need not be drawn. Therefore, only the objects A and B that are determined to be in the table state in FIG. 7 are drawn.

As described above, the occlusion culling determination result can be estimated to be correct for several cycles as long as the amount of movement of the object is small. Therefore, the occlusion culling determination in step S4 of FIG. 5 is performed once in the N loop, where N is a given positive integer. In this case, the drawing with the virtual color needs to be performed only once in the N loop, so that the processing efficiency is improved as compared with the case where this is performed every time.

Further, in the virtual color drawing of FIG. 7, the feature of the present embodiment is that the LOD (level 1) of the object is set lower than the actual LOD (level 0) shown in FIGS. Is. The reason is that drawing with virtual colors is intended only for determining occlusion culling, and it is not necessary to draw using level-0 fine polygons.

Therefore, the drawing apparatus performs processing by combining polygon reduction and occlusion culling as described above. As a result, the speed of occlusion culling determination is increased, and the drawing speed of the object is improved as a whole.

In FIGS. 7, 8 and 9, each sofa is treated as one object, but when one sofa is composed of a plurality of parts, each part is treated as one object, Similar occlusion culling is done.

10 and 11 are more detailed flowcharts of the drawing processing by the drawing apparatus of FIG. Here, it is assumed that the graphics processing device 25 draws using a double buffer as a frame buffer.

First, the CPU 23 substitutes 1 into the control variable I representing the number of loops (step S11 in FIG. 10),
The data of the geometry primitives forming each object is input to the shape data / assembly data memory 31 (step S12).

Next, the vertices of those geometry primitives are subjected to coordinate conversion, the coordinate values viewed from the absolute coordinate system are calculated, and stored in the coordinate conversion data memory 32 (step S13). Then, in steps S14 to S20, a loop process of the normal drawing process (true-color display) is performed, and once every N times this process is repeated, the virtual color drawing process (virtual-color display) of FIG. 11 is performed once. I do.

Here, first, the remainder when I is divided by N is obtained, and it is checked whether or not it is 0 (step S14). If the remainder is 0, orthographic projection or perspective projection projection transformation is performed (step S21 in FIG. 11), and virtual color information of each object is stored in the frame buffer (step S22). This causes the virtual color of each geometry primitive to be stored in the frame buffer along with the pixel information.

At this time, the graphics processing device 25
Represents the screen on the xy plane and the depth direction on the z-axis,
The anteroposterior relationship of points overlapping in the z direction (points having the same xy coordinate values) is determined. This judgment is well known in the drawing process by the existing graphics library, and normally,
It is done at high speed by hardware. Then, the color of the closest point is stored as the color of the corresponding pixel.

The drawing with the virtual color is usually performed by a model in which the LOD is set to a low level, and no special shading processing is performed. Therefore, it is drawn in a so-called solid state.

In the preprocessing of the drawing processing, rgb
A combination of three integers of values (0 to 255, respectively) is assigned to each object in the dictionary order as virtual color identification information. With this, a maximum of 16,777,
Different virtual colors of 216 (= 256 × 256 × 256) colors can be assigned to the object. The correspondence relationship between the identification information of these virtual colors and the objects is stored in the array table memory 36.

Next, the CPU 23 reads the virtual color information of each pixel of the frame buffer into the virtual color information memory 33 (step S23), scans the pixel in the virtual color information memory 33, and appears on the pixel. Extract the virtual color identification information. Then, the array table memory 36 is referred to based on the identification information, and the flags of the objects having these virtual colors are set to 1 in the occlusion culling memory 34 (step S24). Also, the flags of other objects are set to 0 (step S25).

Next, the frame buffer in which the virtual color information has been written is completely cleared once, and swapping of the two buffers is skipped (step S26). Therefore, the image in virtual color is not displayed, and on the screen,
The image drawn in the previous loop will remain.

Next, lighting / shading is set (step S15 in FIG. 10), and projective transformation of orthographic projection or perspective projection is performed (step S1).
6). Then, the flags of the respective objects are compared, and the true color and the pixel information of the geometry primitive are stored in the frame buffer only for the objects whose value is 1 (step S17).

Next, the graphics processing device 25
The two buffers are swapped (step S18), and the accumulated information is displayed on the screen (step S19). In the drawing with the true color, a model having a higher level LOD than that in the drawing with the virtual color in step S22 is used.

Then, setting I = I + 1, step S
The processing after 12 is repeated. Thus, the occlusion culling determination is performed once each time N frames are displayed, and the next N frames are displayed based on the determination result.

Next, the standard of polygon reduction for determining the number of polygons of the model used for drawing in steps S17 and S24 will be described. As the basic algorithm of polygon reduction, for example, the above-mentioned algorithm using the secondary error metric can be used.

At this time, it is desirable to reduce the number of polygons of each object constituting the assembly model while keeping the shape of the entire assembly model as much as possible, and effectively reduce the total number of polygons of the assembly model as a result. For example, the assembly model B0 in FIG.
It is composed of 1, B2, B3, B4, B5 and B6, and it is required to efficiently reduce the number of polygons of the assembly model B0.

Therefore, in FIG. 12, the initial polygon number of each object Bi (i = 1, ..., 6) is Pi, and the radius of the sphere including the object Bi is Ri,
With the radius of the sphere including the assembly model B0 as R0, the target number of reduced polygons Qi is determined by the following equation. Qi = α × Pi × (Ri / R0) (1) where α is a real number parameter satisfying 0 ≦ α ≦ 1,
It is set smaller as the LOD becomes lower. (1)
According to the formula, the target polygon number Qi is the object Bi
It is determined based on the ratio of the radius of the inclusion sphere of B to the radius of the inclusion sphere of the assembly model B0. In other words, the target polygon number Qi is determined based on the ratio between the size of the object Bi and the size of the assembly model B0.

Therefore, by calculating the target polygon number using the equation (1), the polygon number of a relatively small and unobtrusive object in the assembly model can be greatly reduced as compared with other objects. it can. Since such a change in the number of polygons of an object does not significantly affect the overall shape, the shape of the assembly model is not significantly impaired, and the number of polygons is effectively reduced.

For example, objects such as screws and bolts are often composed of many polygons for their size. When the assembly model includes objects of such parts, the polygon reduction of these parts is effectively performed by performing polygon reduction according to the equation (1), and the processing speed is increased.

The information of each object in which the number of polygons has been reduced for each LOD is stored in the polygon reduction memory 35 and used when drawing in the frame buffer.

Here, the LOD of the original assembly model before performing polygon reduction is set to level 0,
In the expression (1), α = 1.0 is set, and the LOD of the assembly model in which the number of polygons is reduced is set to level 1. At this time,
By performing the true color drawing in step S17 of FIG. 10 at level 0 and the virtual color drawing in step S24 of FIG. 11 at level 1, occlusion culling is executed at high speed, and efficient 3D drawing is possible. Becomes

More generally, the LOD can be set in multiple stages by preparing a plurality of values of α in the equation (1). For example, level 0, level I, level J (0 <I <
When the three-stage LOD of J) is set, the true color is drawn at level I and the virtual color is drawn at level J when the assembly model is being moved interactively. When the assembly model is stationary, true color drawing is performed at level 0.

By the way, in the equation (1), the target polygon number is determined based on the ratio of the radius of the containing sphere, but instead, the target polygon number is determined based on the ratio of the surface area and the volume of the containing sphere. May be. Further, instead of the inclusion sphere, another figure such as a cube or a rectangular parallelepiped containing an object or an assembly model may be used, and the target polygon number may be determined based on the ratio of their sizes.

Further, when there are indexes indicating the sizes of the object and the assembly model, the target polygon number can be determined based on those indexes. Also,
If the object is composed of other geometry primitives, the target number of geometry primitives can be determined in the same manner.

In the embodiment described above, the drawing process of the assembly model of the mechanical CAD is mainly described, but similar occlusion culling is also performed in other 3DCG fields such as graphics games, animations and walkthroughs. It can be carried out.

Further, in the occlusion culling determination, instead of drawing with a virtual color, unique identification information may be given to each object and a bitmap in which the object is drawn may be generated using the identification information. Also in this case, similarly to the case of the virtual color, it is possible to determine whether or not an object is occluded by another object by scanning the identification information on the bitmap.

[0082]

According to the present invention, in the 3DCG system, the occlusion culling determination is performed on the bitmap using the shape of the object itself, so that the occlusion culling can be performed accurately and at high speed. This is expected to improve the quality of 3DCG.

[Brief description of drawings]

FIG. 1 is a principle diagram of a drawing apparatus of the present invention.

FIG. 2 is a configuration diagram of a drawing device.

FIG. 3 is a configuration diagram of an information processing device.

FIG. 4 is a diagram showing a recording medium.

FIG. 5 is a flowchart of first drawing processing.

FIG. 6 is a diagram showing polygon reduction.

FIG. 7 is a diagram showing occlusion culling.

FIG. 8 is a diagram showing drawing after occlusion culling.

FIG. 9 is a diagram showing movement of an object.

FIG. 10 is a flowchart (part 1) of the second drawing process.
Is.

FIG. 11 is a flowchart (part 2) of the second drawing process.
Is.

FIG. 12 is a diagram showing a sphere containing an object.

[Explanation of symbols]

1 Shape information storage means 2 table means 3 Virtual image storage means 4 Judgment means 5 Occlusion Culling Means 6 Reduction means 7 Display means 11 Shape / assembly information storage unit 12 Position / posture information update unit 13 Occlusion Culling Judgment Section 14 Drawing unit 15 Polygon reduction calculator 21 Data input device 22, 24 interface 23 CPU 25 graphics processor 26 Graphic display 27 RAM 28 storage 31 Shape data / assembly data memory 32 coordinate conversion data memory 33 Virtual color information memory 34 Occlusion Culling Memory 35 Polygon reduction memory 36 Sequence table memory 41 Portable recording medium 42 database

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Yosuke Senda             4-1, Kamiodanaka, Nakahara-ku, Kawasaki-shi, Kanagawa             No. 1 within Fujitsu Limited F-term (reference) 5B046 FA19                 5B080 AA13 BA07 CA01 CA03 FA02                       GA01

Claims (6)

[Claims] 1. A drawing device for displaying a plurality of objects constituting a three-dimensional graphics model, wherein each object is constructed in accordance with a ratio of the size of each object to the size of the three-dimensional graphics model. Reduction means for reducing the number of geometry primitives to be performed, determination means for performing an occlusion determination using an object with a reduced number of geometry primitives, and an object determined not to be occluded among the plurality of objects, the occlusion A drawing device comprising: a display unit that displays using more geometry primitives than the number of geometry primitives used for the determination. 2. A drawing device for displaying a plurality of objects constituting a three-dimensional graphics model, wherein a ratio of a size of a figure including the three-dimensional graphics model to a size of a figure including each object is set. Determine the target number of geometry primitives that compose each object based on the reduction means to reduce the number of geometry primitives of each object to the target number, and the occlusion judgment using the object with the reduced number of geometry primitives. Means for displaying an object determined not to be occluded among the plurality of objects by using more geometry primitives than the number of geometry primitives used for the occlusion determination. Drawing device. 3. A drawing device for displaying a plurality of objects constituting a three-dimensional graphics model, the drawing device being based on a ratio of a radius of a sphere including the three-dimensional graphics model and a radius of a sphere including each object. And a reduction means for determining the target number of geometry primitives constituting each object and reducing the number of geometry primitives of each object to the target number, and a determination means for performing occlusion determination using the object with the reduced number of geometry primitives. And a display unit for displaying an object determined not to be occluded among the plurality of objects by using more geometry primitives than the number of geometry primitives used for the occlusion determination. apparatus. 4. A recording medium recording a program for a computer that displays a plurality of objects constituting a three-dimensional graphics model, the ratio of the size of each object to the size of the three-dimensional graphics model. In accordance with, reduce the number of geometry primitives that make up each object, perform occlusion determination using the object that has reduced the number of geometry primitives, the object that is determined not occluded among the plurality of objects, A computer-readable recording medium recording a program for causing the computer to execute a process of displaying using more geometry primitives than the number of geometry primitives used for the occlusion determination. 5. A recording medium recording a program for a computer that displays a plurality of objects constituting a three-dimensional graphics model, the size of a figure including the three-dimensional graphics model and each object. Determine the target number of geometry primitives that compose each object based on the size ratio of the figure to be reduced, reduce the number of geometry primitives of each object to the target number, and use the object with the reduced number of geometry primitives. Occlusion determination is performed, and the computer is caused to execute a process of displaying an object determined not to be occluded among the plurality of objects by using more geometry primitives than the number of geometry primitives used for the occlusion determination. A computer-readable recording medium storing a program for. 6. A recording medium recording a program for a computer that displays a plurality of objects constituting a three-dimensional graphics model, the radius of a sphere including the three-dimensional graphics model, and each object. The target number of geometry primitives that compose each object is determined based on the ratio of the radius of the sphere, the number of geometry primitives of each object is reduced to the target number, and the occlusion is performed using the object with the reduced number of geometry primitives. Performing a determination, for causing the computer to perform a process of displaying an object determined not to be occluded among the plurality of objects by using more geometry primitives than the number of geometry primitives used in the occlusion determination. A computer-readable recording medium programs.