JPH11242753A - Method and device for three-dimensional plotting - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically provide out-of-focus display by repeating processing which equally divides luminance value to neighborhood pixels corresponding to a copy position calculated in accordance with the deviation between the Z value of a pixel and the depth of field. SOLUTION: The coordinate and luminance of a pixel are received, the luminance is made into 1/4, the absolute value D1 of finite difference between Zf value of a corresponding image which is preliminarily set to a depth of field register and the Z value of the pixel is calculated, copy pixel distance D2 having a non-linear characteristic to deviation D1 is calculated by using a function of arctan and a result is made an integer (S101 to 103). The coordinates of neighborhood pixels are calculated according to the coordinate and the D2 and if the vertex accumulation bit of a pixel held on a frame memory being 1 is decided (S104 and 105). The luminance of a received pixel is added to the luminance of the neighborhood pixels and luminance value is written to a copy destination pixel (S108). Furthermore, the processing is continued until the processing of all of the neighborhood pixels is finished to the received pixel (S109).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a three-dimensional drawing apparatus, and more particularly to a drawing method for easily and rapidly processing a defocus of a drawing figure.

[0002]

2. Description of the Related Art Conventionally, a process of focusing or blurring a three-dimensional figure is performed by drawing the same figure slightly shifted in position on a frame memory by a user's instruction a plurality of times.
This is realized by adding the image data on another memory (accumulation memory) and finally copying the result on the frame memory. In this example,
"OpenGL Programming Guide (OpenGL ARB, ISB
N4-7952-9645-6c3055), the scene unchairing method is known.

[0003]

In the above prior art, 3
In the case of displaying a three-dimensional figure whose focus is to be blurred when displaying a three-dimensional figure, the user needs to perform drawing a plurality of times while shifting the drawing position, which is troublesome and time consuming.
It has been difficult to apply to real-time GC and the like.

An object of the present invention is to automatically realize a display in which a three-dimensional figure is defocused without depending on a user instruction.
An object of the present invention is to provide a method and an apparatus for drawing a three-dimensional figure.

[0005]

An object of the present invention is to display a three-dimensional graphic on a two-dimensional screen by using coordinate data including depth information (Z value) and luminance value for each pixel in the graphic. In a three-dimensional rendering method of generating display data and rendering the same in a frame memory, the Z value of the target pixel and the depth of field (Zf) set for each screen are sequentially determined for each pixel in the target graphic. The drawing process of obtaining the copy position (D2) of the pixel in accordance with the deviation (D1 = Z-Zf), and dividing the luminance value of the pixel into neighboring pixels corresponding to the copy position is repeated, and the periphery of the target graphic is repeated. This is achieved by forming a blur area (low-luminance area) that increases or decreases according to the deviation.

Alternatively, for each pixel in the target graphic, its Z
Value and depth of field (Zf) set for each screen
And calculates a copy pixel distance (D2) of the pixel based on the deviation (D1) of the pixel, and equally divides the luminance value of the pixel into a plurality of neighboring pixels separated from the pixel by the copy pixel distance in a predetermined direction. This is achieved by writing and, if a luminance value has already been written to a neighboring pixel, performing drawing processing so as to add it to the current value, and repeating for all pixels in the target graphic.

[0007] The copy pixel distance is determined so that the absolute value of the deviation is nonlinearly mapped to an integer space of 0 to n. The predetermined directions are four directions of up, down, left, and right, and the luminance value of this division for writing to the neighboring pixels is １ ／ of the luminance value of the target pixel. Incidentally, in the embodiment, n = 3, and the neighboring pixels on the four sides are located at a maximum of three pixels away from the pixel to be processed.

According to the present invention, when the graphic to be drawn is forward or rearward with respect to the depth of field (Zf) set by the user, a low-luminance blur area is formed outside the periphery of the graphic, The width (the number of pixels) of the area is different from Zf (D
It can be increased or decreased according to 1).

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. FIG. 2 shows a configuration of a computer system to which the present invention is applied. The CPU 202 executes the programs stored in the memory 203, and simultaneously executes the programs stored in the memory 203.
By referring to and updating the coordinate data of the three-dimensional figure, a drawing processing command is issued to the drawing processing apparatus 204 via the data bus 201, and the coordinate data of the three-dimensional figure is transferred.

The rendering processor 204 of this embodiment is characterized by a rendering processor 206. First, the three-dimensional coordinate data received by the geometry processing unit 205 is converted into a two-dimensional coordinate system, and a luminance calculation is performed. Next, the rendering processing unit 206 calculates a span pixel from the vertex data of the figure developed on the two-dimensional coordinate system, and further performs a Z comparison and a blurring process described later for each pixel. Write the pixel value. Frame memory 20
7 is stored in the digital / analog converter 208.
Is periodically converted into an analog signal, and finally displayed on a display.

FIG. 3 shows a configuration of a rendering processing unit according to one embodiment. The rendering processing unit 206 includes a CPU
The apparatus includes a depth-of-field register 301 set from 202, a span pixel calculation processing unit 302, and a luminance correction pixel copy / write processing unit 303.

Here, the depth of field refers to a plane perpendicular to the Z-axis at which a perfectly in-focus object exists within a visual volume formed on a three-dimensional viewpoint coordinate system. I do. Generally speaking, the depth of focus (corresponding to the focus position of the camera) viewed in the depth direction of the screen is usually set by the CPU 202 for each screen.

FIG. 4 is an explanatory diagram showing a data structure per pixel of the frame memory. RGB buffer 401
Occupies 24 bits per pixel, and holds the values of red (R), green (G), and blue (B) with 8 bits each. The Z buffer 402 stores the depth information (z value) as 2
Holds in 4 bits. The vertex accumulation bit is composed of one bit per pixel (AccB) and is initially reset to 0, and the luminance correction pixel copy / write processing unit 203
Used as a processing flag.

Next, the processing of the span pixel calculation processing unit 302 will be described. FIG. 5 is a schematic diagram showing coordinate data and depth of field on a three-dimensional apparatus coordinate system. X axis 501, Y
The axis 502 and the Z-axis 503 constitute a three-dimensional device coordinate system, and the geometry processing unit 2
The coordinate data of the vertices 505, 506, and 507 output by 05 and the depth of field 504 are represented on a three-dimensional apparatus coordinate system. The plane 515 corresponds to the display surface of the display 209. The depth of field 504 is the depth of field register 30
The distance from the origin 500 by the distance of Zf set to 1
A plane perpendicular to the axis 503. The vertices 508, 509 and 510 are obtained by orthogonally projecting the vertices 505, 506 and 507 onto the plane 515, respectively.

The span pixel calculation processing unit 302 receives the vertices 505, 506 and 507 after the three-dimensional geometry processing from the geometry processing unit 205, and outputs the vertices 508 and 509.
And 510, the luminance and the Z value of the pixel existing inside the figure are calculated for each pixel in the same manner as in the related art.

As shown in FIG. 6, (a) the coordinates and the luminance of the pixels forming the line connecting the vertices are calculated by the DDA processing, and (b) the luminance in the raster (X-axis direction) direction is calculated by the rasterizing processing. Is calculated. Then, the calculation result of each pixel is sequentially transferred to the brightness correction pixel copy / write processing unit 303 with its coordinates and brightness. After the above processing is performed for all the pixels inside the figure, the process ends. It should be noted that the dashed-line grid shown in the figure indicates pixels on the display screen 515.

FIG. 7 is a conceptual diagram of the luminance correction pixel copy / write processing. Upon receiving the data of the pixel 514, the luminance correction pixel copy / write processing unit 303 multiplies the luminance by ４, and applies the same に to the pixel 514 and the four neighboring pixels 701 to 704, which are equidistant to the four sides. Write twice the brightness. The coordinates of the received pixel 514 and the coordinate distance 705 between the neighboring pixels are referred to as a luminance correction pixel copy distance (copy pixel distance). Further, when writing the luminance of each pixel, the vertex accumulation bit (AccB) 403 held on the frame memory 207 is determined, and AccB = 1.
Is set, a value obtained by adding the luminance of the pixel before writing the luminance and the current luminance is written.

The above processing is performed by the span pixel calculation processing unit 302.
Is performed on all the pixels calculated, the luminance of the pixel in the central part of the graphic is increased and approaches the original luminance, and the luminance of the pixel in the peripheral part of the graphic becomes close to １ ／ of the original luminance. Thus, the display of the defocused figure is realized.

FIG. 1 shows a flow of a luminance correction pixel copy / write process according to one embodiment. This process is executed each time a pixel inside the figure calculated by the span pixel calculation processing unit 302 is received.

For example, the coordinates (X1, Y1,
Z1) and the brightness (R1, G1, B1), the brightness is multiplied by 1/4 (R1 / 4, G1 / 4, B1 / 4) (S101).

Next, the depth of field register 3 is set in advance by a three-dimensional graphic display program driven in the CPU 202.
The absolute value D1 of the difference between the Zf value of the corresponding screen set to 01 and the Z value (= Z1) of the pixel 514 is calculated by Equation 1 (S102).

[0022]

D1 = | Z1-Zf | This D1 will be described with reference to FIG. 5. The correspondence between the coordinate position 512 on the original figure (vertex 505 to 507) corresponding to the pixel and the depth of field 504 The distance 515 from the position 513 is obtained.

Next, a copy pixel distance D2 having a non-linear characteristic with respect to the deviation D1 is calculated using an arctan function, and the result is converted into an integer (S103).

[0024]

D2 = | tan ~ ¹ (D1) | · 6 / π Here, the value of tan ~ ^{1 is} a value of 2π to -2π, and D2
Takes a value of 0 to 3.

The copy pixel distance D2 according to Equation 2 is an example. Equation 2 non-linearly maps the deviation D1 between the depth of field Zf, which can take a sufficiently large value, and the Z value of each pixel in the figure into an integer space of 0 to 3, and calculates the large deviation of the pixel in the XY axis direction. This is designed so that the neighborhood does not diverge, and can be replaced by another nonlinear function such as sigmoid.

The subsequent processing is repeated for one pixel, and the four neighboring pixels 701 (X1, Y1 + D2) and 702 (X1-X1) of the pixel 514 (X1, Y1) shown in FIG.
D2, Y1), 703 (X1, Y1-D2), 704
This is performed for all of (X1 + D2, Y1).

First, the coordinates (X
From (1, Y1) and D2, the coordinates 701 of the neighboring pixels are calculated (S104). Next, it is determined whether the vertex accumulation bit (AccB) of the pixel 701 held in the frame memory 207 is 1 (S105). Initially, it is "NO" because it has been reset to zero. As a result, AccB = 1 is set for the pixel (S
107). Next, the copy destination pixel (in this case, pixel 70
1 pixel) to the luminance value (R1 / 4, G1 / 4, B1 /
4) is written (S108). Further, until the processing of all the neighboring pixels for the received pixel is completed (S10
9), repeat the processing from S104.

On the other hand, in the determination of S105, AccB = 1 ("
In the case of "YES"), the luminance value has been written once as the neighboring pixel. Therefore, the luminance (R0, G
0, B0), and adds the luminance (R1 / 4, G1 / 4, B1 / 4) of the pixel received this time (R0 + R0).
1/4, G0 + G1 / 4, B0 + B1 / 4) (S1
06).

Thus, in the luminance value correction pixel copy processing of another pixel, if the neighboring pixel matches the copy pixel distance D2, the luminance of another pixel / 4 is added. In general, since the luminance in a figure has continuity, even if the luminance inside the figure is once divided into four parts, the luminance inside the figure is compensated for from other surrounding pixels, and almost the original luminance is recovered. However, at the periphery of a figure having no pixels on the outside, the brightness is reduced and the figure becomes blurred because of insufficient filling.

When the Z value of the received pixel is the same (or approximate) as the depth of field Zf, D2 = 0, so that the luminance value is written to the frame memory as the copy destination = received pixel. You may.

FIGS. 8A and 8B show a display graphic by the blurring processing of the present embodiment, and FIG. 8A shows a graphic 801 on the depth of field,
(B) shows the figure 802 located behind the depth of field. In the case of the graphic 801, since the luminance correction copy pixel distance D 2 = 0, the copying of all the pixels is performed at the original pixel position, and is displayed as a graphic with no defocus. On the other hand, in the case of the graphic 802, the luminance correction copy pixel distance D2 = 1, and the luminance value of the outline of the graphic is reduced, and the graphic is displayed as a defocused graphic.

[0032]

According to the present invention, drawing of a three-dimensional figure is performed by changing the drawing position and brightness of the pixel in accordance with the deviation between the depth of field Zf set for each screen and the Z value of each pixel in the figure. , It is possible to realize a blur display that lowers the luminance of the peripheral portion of the figure.

According to this, since it is not necessary for the user to give an instruction for positional deviation of the same figure, it is possible to provide a three-dimensional figure drawing processing apparatus which is quick and easy to use, and is applicable to real-time CG and the like.

[Brief description of the drawings]

FIG. 1 is a flowchart showing a blur display drawing process in a three-dimensional figure drawing method according to one embodiment of the present invention.

FIG. 2 is a system configuration diagram of a drawing processing apparatus to which the present invention is applied.

FIG. 3 is a configuration diagram of a rendering processing unit according to one embodiment.

FIG. 4 is an explanatory diagram showing a data configuration per pixel of a frame memory.

FIG. 5 is a schematic diagram showing coordinate data and a depth of field on a three-dimensional apparatus coordinate system.

FIG. 6 is an explanatory diagram of a DDA process and a rasterizing process for calculating a luminance and a Z value of a pixel existing inside a figure by a span pixel calculation processing unit.

FIG. 7 is a conceptual diagram of a luminance correction pixel copy / write process.

FIG. 8 is an explanatory diagram showing a display figure of a drawing process according to the embodiment.

[Explanation of symbols]

201: data bus, 202: CPU, 203: memory, 204: drawing processing unit, 205: geometry processing unit, 206: rendering processing unit, 207: frame memory, 208: digital / analog converter, 209 ...
Display, 301 ... depth of field register, 302 ...
Span pixel calculation processing unit 303: luminance correction pixel copy / write processing unit

Claims (5)

[Claims] When displaying a three-dimensional graphic on a two-dimensional screen, display data consisting of coordinate data including depth information (Z value) and a luminance value is generated for each pixel in the graphic, and a frame memory is generated. In the three-dimensional drawing method, the pixel is sequentially copied for each pixel in the target figure in accordance with the deviation between the Z value of the target pixel and the depth of field (Zf) set in advance in screen units or the like. A drawing process of obtaining a position and dividing the luminance value of the pixel into a neighboring pixel corresponding to the copy position is repeated, and a blur region which increases or decreases according to the deviation is formed at a peripheral portion of the target graphic. 3D drawing method. 2. When displaying a three-dimensional graphic on a secondary screen, display data including luminance data and coordinate data including depth information (Z value) is generated for each pixel in the graphic and drawn on a frame memory. In the three-dimensional drawing method, a copy pixel distance of a pixel in a target graphic is calculated based on a deviation between a Z value and a depth of field (Zf) set in advance on a screen basis or the like. The luminance value of the pixel is equally divided and written into a plurality of neighboring pixels separated by the copy pixel distance in a predetermined direction from the pixel, and if the neighboring pixel has already been written with the luminance value, the current value is added. A three-dimensional rendering method, wherein the rendering process is repeated for all pixels in the target graphic. 3. The three-dimensional drawing method according to claim 2, wherein the copy pixel distance is obtained such that the absolute value of the deviation is nonlinearly mapped to an integer space from 0 to n. 4. The method according to claim 2, wherein the predetermined direction is four directions of up, down, left, and right, and a luminance value of the division for writing in the neighboring pixel is 1/1/1 of a luminance value of a target pixel. 4. A three-dimensional drawing method, characterized in that: 5. A C for issuing a drawing processing command for a three-dimensional figure
A PU and a drawing processing device that converts the three-dimensional coordinate data into a two-dimensional coordinate system based on the received drawing processing command, calculates a luminance value, and writes a pixel value to a frame memory; The drawing processing device stores a depth-of-field register that stores a depth-of-field (Zf) set for each screen from the CPU, and stores coordinate data including a Z value of a pixel existing inside a graphic to be processed and luminance. A three-dimensional drawing apparatus, comprising: a span pixel calculation processing unit for calculating each pixel; and a luminance correction pixel copy / write processing unit for performing the drawing processing according to claim 1. .