JPH05181950A - Three-dimensional moving image generation system - Google Patents

Abstract

PURPOSE:To obtain a three-dimensional moving image generation system which avoids and substitute the repetition of a process for development into image data on bodies with a faster rectangular transfer process so as to greatly reduce the delay of an image generation time due to the repetition of the same image generating process between the bodies. CONSTITUTION:A still picture component of a stationary body on a screen and moving picture components of individual bodies in parallel motion on the screen are generated in a 1st and a 2nd area of an image memory which do not overlap with each other (S12-S14). Specific depth is added to the respective moving picture components in each screen generation cycle, rectangular transfer to a specific position in the still picture component is performed (S15), and a three-dimensional image is composed by hidden-surface processing (S16). Partial images of the still picture component which is lost as a result of the composition of the moving picture components are saved in a 3rd area of the image memory (S15) and the source still picture component is restored in a next screen generation cycle (S19).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a three-dimensional moving image generation system for generating a three-dimensional moving image using an electronic computer or the like and displaying the generated three-dimensional moving image on a screen or the like.

[0002]

2. Description of the Related Art FIG. 2 is an example of an image showing one frame of a three-dimensional moving image, in which two spherical objects A and B move in an object C composed of a basket container under the action of gravity or the like. It shows how to move and collide.

As shown in FIG. 2, the image includes three objects A,
It is composed of B and C. Object C is stationary. The objects A and B are simulated by a computer for parabolic movement, mutual collision and collision with the inner wall of the object C, their three-dimensional positions are changing every moment, and images visualizing them are generated one after another. And is displayed as a moving image on the CRT (CRT) screen. FIG. 3 shows a schematic functional block diagram of a conventional three-dimensional moving image generating apparatus that generates such a three-dimensional moving image.

The three-dimensional moving image generating apparatus shown in FIG. 3 has a main memory 1 for storing data, and a central processing unit (hereinafter referred to as CPU) 2 for controlling the entire apparatus is connected to the main memory 1. There is. An image memory 3 for storing image data is connected to the CPU 2, and a CRT 5 is connected to an output side of the CPU 2 via a video tape recorder (hereinafter referred to as a video) 4. Further, the CPU 2 is connected to a depth buffer (hereinafter, referred to as Z buffer) 6 that stores the depth Z of the image.

In this type of three-dimensional moving image generating apparatus,
The main memory 1 stores the shape data of the objects A, B, and C to be displayed, and the CPU 2 retrieves the shape data of the objects A, B, and C.
An image of C is generated and written in an appropriate position in the image memory 3. The image memory 3 is a memory that stores at least one screen of image data. In the image memory 3, in an example of a full-color image with a brightness value, each pixel has 6 pixels of red (R), green (G), blue (B), and brightness (I).
Bits, the contents of which are always imaged on the CRT 5 by the video 4. Furthermore, a Z buffer 6 for storing the value of the depth Z for each pixel, for example, 12 bits is provided.

When the image of the object is generated, the CPU 2 obtains the depth Z for each pixel, compares the depth Z with the depth stored in the Z buffer 6, and based on the result, the image data and the depth Z.
Hidden surface processing is performed to determine whether or not to write, but this hidden surface processing is not performed by the CPU 2 but by the Z buffer 6
In many cases, it is automatically performed by a comparison calculation circuit (not shown) attached to the.

FIG. 4 is a flowchart showing a processing procedure of a conventional three-dimensional moving image generation system using the apparatus of FIG. In the conventional three-dimensional moving image generation method, when the operation of the apparatus is started, the positions of the objects A, B, and C are set in process S1, and then the process proceeds to one-screen image generation process S2. In this one-screen image generation processing S2, first, in processing S2-1, the background color and the brightness value (for example, R = G = 0,
B = 63, I = 32) and the maximum depth value (for example, Z = 2047) is written in the entire Z buffer 6.

Next, in step S2-2, the shape data of the object C is fetched from the main memory 1, expanded into image data based on the display position information, and the image memory 3 and the Z buffer 6 are processed.
Written in. For example, R = B = 0, G = 63
Is an image in which I is changed for each surface of the object C and Z is changed for each pixel in the surface. Then, in process S2-3, the object A
However, in the process S2-4, the object B is developed into the image data of the sphere and written in the image memory 3 and the Z buffer 6. This is an image in which, for example, a yellow sphere with R = G = 63 and B = 0 and Z and I change for each pixel.

When the image generation process S2 for one screen is completed, it is determined in process S3 whether or not the next screen is generated. In the case of generation, the next position of the objects A and B is calculated and the process S
Return to 2. In this way, the desired three-dimensional image can be obtained by sequentially performing the expansion processing on the pixels of the objects A, B, and C and combining them by the hidden surface processing.

[0010]

However, in the conventional three-dimensional moving image generation method, since the expansion processing of the objects A, B, and C into images must be performed every time one image is generated. This requires a long processing time, so that the performance as a three-dimensional moving image generation device cannot be fully exhibited. Although the object C is stationary on the screen, the same image is repeatedly generated at the same display position, resulting in duplication. This is because as a result of superimposing the images of the moving objects A and B, a part of the image of the stationary object C is lost and cannot be used next time.

For the moving objects A and B, the relatively time-consuming process of generating a spherical image must be repeated twice in one screen generation cycle. Therefore, an object with such a shape that takes a long time to generate an image
The reality is that it cannot be handled as a component of a real-time moving image. According to the present invention, in the conventional three-dimensional moving image generation method, the delay of the screen generation time caused by repeatedly executing the same image generation process for each object is significantly reduced. The present invention provides a three-dimensional moving image generation method that avoids repetitive execution and substitutes for a faster rectangular transfer process.

[0012]

In order to solve the above-mentioned problems, the present invention provides a three-dimensional moving image generation system using an image memory with a depth buffer, in which a still image component consisting of an object stationary on the screen is generated. A process of generating in the first area of the image memory, a process of generating moving image components of each moving object performing parallel motion on the screen in the second area of the image memory, and each of the moving images. A process of adding a predetermined depth to the component, transferring the rectangle to a predetermined position in the still image component, and synthesizing by a hidden surface process, and changing the position and the depth for each screen generation cycle to generate a three-dimensional moving image. The partial image of the still image component lost by the process of generating and combining the moving image components is stored in the third memory of the image memory.
And the process of restoring the original still image component in the next screen generation cycle.

[0013]

According to the present invention, since the three-dimensional moving image generation system is configured as described above, a still image component consisting of an object that is stationary with respect to the screen and individual motions that perform parallel motion with respect to the screen. The moving image component of the object is generated in advance in the first and second areas of the image memory that do not overlap each other. A predetermined depth is added to each moving image component in each screen generation cycle, the rectangle is transferred to a predetermined position in the still image component, and a three-dimensional image is combined by hidden surface processing. Then, the partial image of the still image component lost due to the combination of the moving image components is saved in the third area of the image memory, and the original still image component is restored in the next screen generation cycle. Therefore, the above problem can be solved.

[0014]

EXAMPLE FIG. 5 is a diagram for explaining a three-dimensional moving image generation system of an example of the present invention on a screen.

This three-dimensional moving image generation method is shown in FIG.
Using a three-dimensional moving image generation device such as
This is a method for generating a three-dimensional moving image. This system requires a sub-screen capable of storing some partial images, in addition to the main screen on which the target moving image is displayed. Figure 5
Then, three rectangular areas P, Q, R of the same size in the sub-screen
, The spherical image is displayed in the area P, and the partial images of the rectangular areas a and b in the main screen are cut out and displayed in the areas Q and R. The main screen and the sub screen are obtained by dividing the image memory 3 and the Z buffer 6 in FIG. 3 by software, but the sub screen portion is not the target image, so it is not always displayed on the CRT 5. May be.

FIG. 1 is a flow chart showing a processing procedure of a three-dimensional moving image generation system according to an embodiment of the present invention. This 3
In the three-dimensional moving image generation method, the moving image of FIG. 2 is generated on the main screen of FIG. 5, but this is performed by a processing procedure different from the conventional method of FIG.

First, when the image generation operation is started, the process S
At 11, the positions of the objects A, B and C are set. Process S12
Then, a sphere image representing the objects A and B is generated in the area P in the sub-screen. In steps S13 and S14, a still image including the background and the object C is generated on the main screen. When entering the moving image generation loop, the rectangular area a on the main screen where the object A is to be displayed is determined in step S15, and the still image cut out by this area boundary is transferred to the rectangular area Q of the sub screen. Similarly, the still image included in the rectangular area b on the main screen where the object B is to be displayed is transferred to the rectangular area R. For example, rectangular transfer a → Q
The processing content of is expressed by the following equation. IM (Xa + i, Ya + j) → IM (XQ + i, YQ + j) Z (Xa + i, Ya + j) → Z (XQ + i, YQ + j) where i = 0, 1, ..., M-1 j = 0, 1, ..., N-1 Here, the starting point of the rectangular area a is (Xa, Ya), the starting point of the rectangular area Q is (XQ, YQ), and the sizes of both rectangular areas are MX.
N, and IM (X, Y) and Z (X, Y) are the image memory 3
And the value at the address (X, Y) of the Z buffer 6 is represented. When the areas a and b have been saved in the areas Q and R in this manner, the image of the sphere in the rectangular area P is transferred to the areas a and b in step S16 and is three-dimensionally combined with the still image. The rectangular transfer P → a executed here is expressed, for example, by the following equation. If Z (XP + i, YP + j) + Za <Z (Xa + i, Ya + j), then IM (XP + i, YP + j) → IM (Xa + i, Ya + j) Z (XP + i, YP + j) + Za → Z (Xa + i, Ya + j) where i = 0, 1, ..., M-1 j = 0,1, ..., N-1 This reads an image in the rectangular area P, adds Za to the depth thereof, and transfers the image to the rectangular area a. Pixel depth that has been stored so far Z (Xa + i, Ya +
j) is the depth Z (XP + i, Y) to be newly written
Only when it is larger than P + j) + Za, the hidden surface processing instruction is given so as to actually perform writing.

In the upper half of FIG. 5, the depth distribution of the sphere stored in the region P and the depth distribution when moving to the regions a and b are drawn, and the depth distribution of this and the still image (background and object C) is shown. Shows the relationship. In the area a, the sphere hides the object C,
It can be seen that in the region b, the sphere has a part that hides the object C and a part that is hidden by C. Further, the depth around the sphere in the rectangular area P needs to be set larger than the background depth of the still image so that it is always hidden. Here, for example, the background depth of a still image is 1000 and the background depth of a sphere is 10.
Since it is set to 23, as long as 0 <Za <1024, the periphery of the sphere is not always displayed. The reason why the background is hidden by the depth data itself is to simplify the algorithm by keeping the region boundaries rectangular. The three-dimensional image thus completed is processed by the CPU 2 in step S18 when the next screen is generated after it is determined in step S17 of FIG. 1 whether or not the next screen is generated.
While calculating the next position of objects A and B, CRT5
Is displayed in. When new positions of the objects A and B are obtained, the previously saved still image is restored by the rectangular transfer Q → a, R → b in step S19 before using the new position to start the next screen generation cycle. The rectangular transfer Q → a is expressed, for example, by the following equation. IM (XQ + i, YQ + j) → IM (Xa + i, Ya + j) Z (XQ + i, YQ + j) → Z (Xa + i, Ya + j) where i = 0,1, ..., M-1 j = 0,1, ..., N-1 Here, not the new position (Xa1, Ya1) of the object A, but the previous position (Xa, Ya) is used as the starting point of the rectangular area of the transfer destination. When the restoration of the still images of the objects A and B is completed, the process returns to the process S15, the next screen generation cycle is started, and the three-dimensional image in which the objects A and B are placed at positions and depths different from the previous time in the processes S15 and S16. Is generated.

By repeating the screen generation cycle through the processing S17 in this way, a moving image in which the objects A and B move around in the object C is generated, and the image generation processing is ended.
This screen generation cycle does not include the process of expanding the shape data from the shape data of the objects A, B, and C to the image, and the rectangular transfer of the part surrounding the objects A and B is replaced, which is extremely fast. Image generation is possible. Therefore,
It is important to execute at high speed the rectangular transfer for saving and restoring that holds the depth used here, and the rectangular transfer for displaying the moving image component accompanied by the hidden surface processing by using the depth.

FIG. 6 shows such a three-dimensional rectangular transfer as a CP.
An example of the configuration of a three-dimensional rectangular transfer circuit that executes faster without using U2 is shown. This three-dimensional rectangular transfer circuit is provided with an image memory input / output buffer 10, a Z buffer data processing unit 20, an address generation unit 30, and a transfer control unit 40 around the image memory 3 and the Z buffer 6. In FIG. 6, the transfer from the rectangular area P to the rectangular area a accompanied by the hidden surface processing is taken as an example, and the necessary parameters are already set in the respective parts by the CPU 2.

The image memory input / output buffer 10 includes an input / output buffer register (hereinafter referred to as RGBI register) 1
1 and has a memory write gate 12 on its input / output side.
And the memory read gate 13 are connected. The output side of the gate 12 and the input side of the gate 13 are connected to the image memory 3
Is connected to the terminal IO. A write gate 14 from the CPU 2 is connected to the input side of the RGBI register 11, and a read gate 15 is connected to the output side of the RGBI register 11.

The Z buffer data processing unit 20 has an input / output buffer register (hereinafter referred to as ACC) 21, and a memory write gate 22 and a memory read gate 23 are connected to the output side thereof. The output side of the gate 22 and the input side of the gate 23 are connected to the terminal IO of the Z buffer 6. A write gate 24 from the CPU 2 is connected to the output side of the gate 23.

On the output side of the ACC 21, a read gate 2 is provided.
5 is connected. Also, on the input side of ACC21,
An output side of an arithmetic logic operation unit (hereinafter, referred to as ALU) 26 is connected, and a flag (hereinafter, ZF) is connected to the ALU 26.
27) is connected. The input side of the gate 22 is connected to the input side of the ALU 26, and the gate 2
The output side of 9 is connected, and the transfer destination depth register (Za) 28 is connected to the input side of the gate 29.

The address generator 30 has a transfer source start point register 31 and a transfer destination start point register 32 connected to the output side of the CPU 2, and a transfer source address counter 33 and a transfer destination address counter 34 are provided on each output side thereof. Are connected. A multiplexer 35 is connected to the output sides of the counters 33 and 34 so that the addresses (X, Y) of the image memory 3 and the Z buffer 6 are given. The transfer control unit 40 has a rectangular size register 41 and a command register 43 connected to the output side of the CPU 2, and a pixel counter 42 is connected to the output side of the register 41.
Each part is controlled so that the rectangular transfer is performed completely once.

Next, the operation of the three-dimensional rectangular transfer circuit shown in FIG. 6 will be described. The start point (XP, YP) of the area P is set in the transfer source counter (XS, YS) 33, and the multiplexer 35
The contents of the image memory 3 and the Z buffer 6 are read through. The read data of the image memory 3 is held in the RGBI register 11. In the Z buffer data processing unit 20, the value Za of the transfer destination depth register 28 is set in the ACC 21 through the gate 29 and the ALU 26 before reading. When the read data from the Z buffer 6 is ZS, ZS enters the ALU 26 through the gate 23. AL
U26 then acts as an adder, resulting in ZA + Z
The value S is set in ACC21.

Next, the start point (Xa, Ya) of the area a is set in the transfer destination address counter (XD, YD) 34, and the multiplexer 35 is used to instruct the image memory 3 and the Z buffer 6. However, the image memory 3 is not read,
Only the Z buffer 6 is read, and the read data is ZD
Then, ZD enters ALU 26 through gate 23. At this time, the ALU 26 acts as a subtractor, and ACC-
ZD subtraction is performed, and the sign bit of the result is ZF2.
Set to 7. At this time, the value of ACC21 does not change,
It remains ZA + ZS. If the sign bit ZF is 0 (that is, ZA + ZS> = ZD), writing to the image memory 3 and the Z buffer 6 is not performed. Sign bit Z
When F is 1 (that is, ZA + ZS <ZD), the value of the RGBI register 11 is written in the image memory 3 and the value of ACC 21 is written in the Z buffer 6.
This matches the above-mentioned hidden surface processing algorithm. In this way, the processing of the first pixel of the rectangular transfer is completed.

Next, if the values of XS and XD are incremented by 1 and the same processing is repeated, the transfer of the second (right adjacent) pixel in the rectangular area is performed. The transfer control unit 40 scans the entire rectangular area using the pixel counter 42, controls so that transfer processing is performed for all pixels, and notifies the CPU 2 when the processing is completed. The CPU 2 only sets parameters for rectangular transfer, starts up and waits for the end, and does not participate in rectangular transfer itself. Therefore, such 3
If the three-dimensional rectangular transfer circuit is provided, the three-dimensional rectangular transfer is executed at an extremely high speed, and the processing method according to this embodiment can be realized in a more effective form.

Next, the CPU 2 shown in FIG.
A method of generating a sphere image in a relatively short time will be described with reference to FIGS. 7 and 8. FIG. 7 is a flowchart showing the processing procedure of the sphere image generation method in the embodiment of the present invention, and FIG. 8 is an explanatory diagram of the sphere image generation method of the embodiment of the present invention. In FIG. 7, first, in step S21, the entire rectangular area is cleared. The background depth of the sphere is, for example, Z = 1023. As shown in FIG. 8, it is assumed that a sphere having a center 0 (XC, YC, α) and a radius α is created on the image memory 3, and the origin is moved to the center 0 of the sphere.
And The depth of each pixel on the spherical surface 50 is obtained by the following double loop of (1) and (2).

(1) In step S22 of FIG. 7, a semi-circular table having a radius α is created at 2 * α consecutive addresses in the main memory 1 of FIG. This shows the shape of one side of the cross section when the sphere is cut along the plane passing through the y axis, as indicated by 51 in FIG. 8, and β is set for each y value in the y section [−α, + α].
= SQRT (α * α-y * y) is stored. In process S23, the value of y is scanned from -α to + α, and in process S24, the value of β with respect to y is obtained by referring to this semicircular table.

(2) In step S25 of FIG. 7, a semi-circular table having a radius β is created at 2 * β consecutive addresses in the main memory 1 of FIG. This shows the shape on one side of the cross section when the sphere is cut along a plane perpendicular to the y-axis, as indicated by 52 in FIG. 8, and z for each x value in the section [−β, + β] of x.
= SQRT (β * β-x * x) is stored. In step S26, the value of x is scanned from −β to + β,
In process S27, the value of z with respect to x is obtained by referring to this semi-circular table.

In the double loop, in step S28, (x,
The values of R, G, B, I, and Z to be written in the image memory 3 and the Z buffer 6 are obtained using (y, z). For example,
R = G = B = 63, I = (− x−y + z) * 32 / α
If a simple calculation such as (negative is set to 0), a white sphere with the brightest front left front is drawn. Moreover, z = α−z is set so that the value of z becomes 0 at the center O. Process S2
At 9, write to address (XC + x, YC + y). In addition,
The semi-circular table is created by an integer addition / subtraction algorithm, and multiplication and square root operation are unnecessary. Even if the above-described efficient processing method is adopted, since the image development of the sphere requires 10 times or more the time of the rectangular transfer of the same size, the rectangular transfer is used.

As described above, in this embodiment, the still image component and the moving image component of the three-dimensional moving image are separately generated, and are combined only by the rectangular transfer in the actual screen generation loop shown in FIG. Therefore, extremely high-speed image generation is possible. Since this rectangular transfer can be executed by using a dedicated three-dimensional rectangular transfer circuit as shown in FIG. 6, for example, the performance can be improved 10 times or more as compared with the case where image generation is repeatedly executed. Further, since a common original image can be used in the generation of a three-dimensional moving image composed of a plurality of moving image components having the same shape, it is particularly suitable for displaying a moving image of a molecular model or an atomic model composed of a large number of particles. The moving image component can be any shape, size,
Even if you give colors and patterns, it is generated in the pre-processing, so
The performance of the video generation loop is not degraded.

The present invention is not limited to the illustrated embodiment, and for example, the contents of the flowcharts of FIGS. 1 and 7 may be changed to other contents, or the circuit configurations of FIGS. 3 and 6 may be changed to other contents. Various modifications are possible, such as changing the configuration.

[0034]

As described above in detail, according to the present invention, a still image component and a moving image component of a three-dimensional moving image are separately generated and stored in the first and second areas of the image memory. I'll do it. Then, in the image generation loop, since the combination is performed only by the rectangular transfer, it is possible to generate an image at an extremely high speed. Since this rectangular transfer can be executed by using, for example, a three-dimensional rectangular transfer circuit or the like, it is possible to further improve the performance as compared with the case where the image generation is repeatedly executed.

In addition, since a common original image can be used in the generation of a three-dimensional moving image composed of a plurality of moving image components of the same shape, various kinds such as displaying a moving image such as a molecular model or an atomic model composed of many particles can be used. It can be applied to the display of moving images. Moreover, the moving image component can be any shape, size, color,
Even if a pattern and a pattern are given, they are generated by the preprocessing, so that the performance of the moving image generation loop is not deteriorated. Therefore, it is possible to provide a high-performance three-dimensional moving image generation method capable of high-speed processing.

[Brief description of drawings]

FIG. 1 is a flowchart showing a processing procedure of a three-dimensional moving image generation system according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a three-dimensional moving image.

FIG. 3 is a functional block diagram showing a configuration of a conventional three-dimensional moving image generation device.

FIG. 4 is a schematic diagram of a conventional 3 using the 3D moving image generating apparatus of FIG.
It is a flow chart which shows a processing procedure of a three-dimensional moving picture generation system.

FIG. 5 is an explanatory diagram of a three-dimensional moving image generation system showing an embodiment of the present invention.

FIG. 6 is a diagram showing a configuration example of a three-dimensional rectangular transfer circuit according to an embodiment of the present invention.

FIG. 7 is a flowchart showing a processing procedure of a sphere image generation system according to an embodiment of the present invention.

FIG. 8 is an explanatory diagram showing a sphere image generation system according to an embodiment of the present invention.

[Explanation of symbols]

1 Main Memory 2 CPU 3 Image Memory 4 Video 5 CRT 6 Depth Buffer S11 Position Setting Processing for Objects A, B, C S12 Processing for Generating Images of Objects A, B in Rectangular Area P of Sub-screen S13 Main Background Color Processing to clear by S14 Image generation processing of still object C S15 Still image saving processing S16 Moving image component combining processing S17 Judgment processing S18 Next position calculation processing of objects A and B S19 Still image restoration processing

Claims (1)

[Claims] 1. A three-dimensional moving image generation method using an image memory with a depth buffer, a process of generating a still image component consisting of an object stationary on the screen in a first region of the image memory, and the screen. Generating a moving image component of each moving object performing parallel motion in the second area of the image memory, and adding a predetermined depth to each moving image component to give a predetermined value in the still image component. Process for transferring a rectangle to the position and combining by hidden surface processing, a process for generating a three-dimensional moving image by changing the position and depth for each screen generation cycle, and a still image lost by combining the moving image components. A three-dimensional moving image generation, characterized in that the partial image of the component is saved in the third area of the image memory and the original still image component is restored in the next screen generation cycle. method.