JPH11161819A - Image processing apparatus, image processing method, and recording medium recording image processing program - Google Patents

Abstract

(57) [Summary] To provide an image processing capable of efficiently performing a perspective interpolation operation. The present invention does not calculate pixel parameters by interpolation processing of polygon parameters obtained by perspective transformation using a perspective value at an internal division ratio on screen coordinates as in the conventional example. An internal division ratio on the world coordinates is once obtained from the perspective value and the internal division ratio on the screen coordinates and recorded. Then, an interpolation calculation process is performed on the parameters of the polygon based on the internal division ratio on the world coordinate to obtain the parameters of the pixel. In such a perspective interpolation operation, a parameter of a pixel on a world coordinate can be directly interpolated. As a result, the parameters of the polygons on the world coordinates can be directly interpolated, so that the conventional step of dividing by the perspective value in the perspective interpolation operation is eliminated. Further, since the internal division ratio on the world coordinates can be calculated only in accordance with the accuracy in the screen screen, the arithmetic processing is light.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus using a computer, and more particularly to an image processing apparatus for interpolating pixel parameters from polygon drawing parameters.
The present invention relates to an image processing apparatus capable of efficiently performing perspective interpolation (hereinafter, perspective interpolation) in consideration of a perspective value in a perspective space, a method thereof, and a recording medium storing the image processing program.

[0002]

2. Description of the Related Art An image processing technique using a computer is used in a simulation device, a game device, or the like. Normally, polygon data on screen coordinates to be drawn on the display screen is obtained from polygon data on world coordinates generated by a simulation or game program, image data for each pixel in the polygon is obtained, and the image data is obtained. Store in the frame buffer memory. Then, an image is displayed on the display device according to the image data in the frame buffer memory. The process of obtaining polygon data on the screen coordinates described above is usually performed by
The processing performed by the geometry processing unit to obtain image data for each pixel from the polygon data is performed by the renring processing unit.

The geometry processing unit converts polygon data on three-dimensional world coordinates into polygon data on two-dimensional screen coordinates. This transformation is generally a perspective transformation to give a perspective. That is, a polygon in world coordinates is projected onto a display screen located at a predetermined distance from the viewpoint with the viewpoint as the center by perspective transformation, and polygon data on screen coordinates is obtained.

On the other hand, in computer graphics using polygons, parameters for giving a texture to a polygon are provided for each vertex of the polygon. The parameters include, for example, texture coordinates, transparency, luminance, color, normal vector, and the like. Therefore, in the rendering process, the parameters of the pixels to be drawn are obtained by performing an interpolation operation on the parameters given for each vertex. However,
In this interpolation calculation, it is necessary to perform not a simple interpolation calculation but a perspective interpolation in consideration of a perspective value in order not to impair the perspective.

FIG. 8 is a view for explaining perspective projection, which is the principle of perspective transformation. In the figure, the world coordinates are X, Y, Z
, And the screen coordinates are indicated by sx and sy. FIG. 8 shows an example in which a linear polygon 10 placed in three-dimensional world coordinates and perpendicular to the X axis is perspective-transformed into two-dimensional screen coordinates. In this perspective transformation, the screen 12 is placed at a distance d from the viewpoint V in the Z-axis direction,
This is an operation for obtaining screen coordinates when the polygon 10 in world coordinates is projected on the screen 12.

Assuming that a straight polygon 10 has end points A and B and a middle point C as shown in the figure, the perspective transformation
From the viewpoint V to each point A, B, C with a straight line,
The positions where the straight line crosses the screen are projected points As, Bs, and Cs on the screen. For example, if the world coordinates of the point A are (xa, ya, za), since the triangles VAA ₀ and VA _s S are similar, the y coordinate of the point As on the screen is sya = ya × d / Za
Becomes Similarly, the screen coordinates for point B are sy
b = yb × d / Zb, the screen coordinates for point C are
syc = yc × d / Zc. The point C is a middle point between the points A and B in the world coordinates, but the projected point Cs of the point C on the screen 12 is the projected point Ca, C as shown in the figure.
The position is shifted to the projection point Bs side from the middle point of b. This produces the perspective created by perspective transformation by perspective projection. That is, d / z generated from the Z value and the distance d to the screen is a perspective value. By using this perspective value, it is possible to obtain a position on screen coordinates with a sense of perspective.

The distance d from the viewpoint V to the position of the screen is always set to 1, and the perspective value is calculated as 1 / z. Finally, the size of the screen is reduced or enlarged according to the distance from the viewpoint V to the screen. (D times), it is also possible to perform an operation with a sense of perspective.

As described above, parameters for giving a texture to a polygon are given for each vertex. For example, in the example of FIG. 8, it is assumed that vertex A is provided with red color data and vertex B is provided with green color data. In this case, the color data at the midpoint C is obtained by interpolation of the color data at the vertices A and B on both sides. That is, when the interpolation is performed based on the positional relationship in the world coordinates, the color data at the midpoint C is 50
% And green are 50%. However, when the interpolation is performed based on the positional relationship in the screen coordinates, the point Cs is located closer to the point Bs, so that the color data at the point Cs has less than 50% of red and more than 50% of green, and impairs perspective. I will be. Since the rendering process is performed on a pixel that is a display point on the screen, it is necessary to perform an interpolation operation based on a positional relationship within screen coordinates.

As means for solving such a problem, a perspective interpolation method has been proposed, for example, in Japanese Patent Laid-Open No. Hei 4-220783. That is, in the perspective interpolation method, the parameters (x, y, z, I) of each vertex of the polygon are set in order to enable linear interpolation on screen coordinates.
Is converted to screen coordinates, a linear interpolation operation is performed using the converted parameters, and then the parameter I is inversely converted again. Conversion to screen coordinates is performed by multiplying the above-mentioned perspective value d / z. By multiplying the color data parameter I by the perspective value d / z, a perspective can be added, so that the linear interpolation operation can be performed according to the positional relationship on the screen coordinates. The inverse transformation is performed by dividing by the perspective value d / z.

FIG. 9 is a flowchart of image processing using the above-mentioned conventional perspective interpolation.
FIG. 4 is a diagram for explaining the perspective interpolation. As shown in FIG. 10A, an example in which a triangular polygon 14 is rendered will be described. FIG. 10A shows a polygon in world coordinates. Polygon 14 is
It has vertices A, B, and C. The vertex A has a coordinate value (x ₀ , y ₀ , z ₀ ) on the world coordinates and an associated parameter I ₀
And similarly, at vertex B, (x ₁ , y ₁ , z ₁ )
Parameter I ₁ and the accompanying is, the vertex C (x _2,
y ₂ , z ₂ ) and the associated parameter I ₂ are given, respectively. Data of these polygons is generated (S1
0).

From the polygon data, the geometry
The above-described perspective transformation is performed by the transformation unit, and the transformation is performed as shown in FIG.
The data on the screen coordinates to be calculated is obtained. Immediately
For the vertex A, the coordinate value on the screen coordinates (s
x₀(= X₀d / z₀), Sy ₀(= Y₀d /
z₀)), Perspective (d / z)₀And accompanying parameters
(Id / z)₀Are required respectively. Vertices B and C
Similarly, as shown in FIG.
(S11).

Then, as shown in FIG. 10 (3), the screen coordinate value (sx) of the left edge point Ls on the side AsBs
_l , sy), the perspective value (d / z) _l , and the associated parameter (Id / z) _l are obtained by linear interpolation according to the positional relationship on the screen. Also, the side As
The same applies to the right edge point Rs on Cs (S12). Parameter I for generating image data
Since is multiplied by the perspective value d / z, it can be obtained by a linear interpolation operation using the internal division ratio t = (sy−sy ₀ ) / (sy ₁ −sy ₀ ) based on the positional relationship on the screen.

Next, as shown in FIG. 10D, the screen coordinate value (sx, sy) and the perspective value (d / d) at the pixel PX on the horizontal scanning line between the two edge points Ls, Rs.
z) _p and the associated parameter (Id / z) _p are respectively obtained by linear interpolation according to the positional relationship on the screen (S13). The interpolation operation is also determined by using the internal ratio according to a positional relationship on the screen _{t = (sx-sx l)} / (sx r over sx _l).

Finally, the parameter (Id / z) _p of the pixel PX is divided by the perspective value (d / z) _p to obtain an appropriate parameter _Ip (S1).
4). The parameters I _p of all the pixels in the polygon are calculated by the above raster scan method, and finally image data is calculated from the parameters I _p and written into the frame buffer (S15), and the image data is stored in the display device. It is supplied and displayed (S16).

According to the above-described perspective interpolation method,
Since the interpolation calculation can be performed based on the positional relationship on the screen coordinates, the interpolation calculation can be performed according to the pixel position on the screen, which is convenient for image processing.

[0016]

However, in the above-described perspective interpolation method, the division shown in step S14 in FIG. 9 must be performed for every pixel every time. In general, division for computers is
This is a heavy processing requiring the most man-hours among the four arithmetic operations. Therefore, the conventional perspective interpolation operation in which this division is required every time in the interpolation operation for all the pixels,
The image processing time is lengthened to make it difficult to perform real-time processing required in games, simulations, and the like. Furthermore,
As the number of parameters increases, the number of times of the above-described division also increases, thereby increasing the image processing time. In addition, when the conventional perspective interpolation calculation is used, the size of the image processing apparatus is increased.

Second, steps S11, S12, and S12 in FIG.
In S13, there is an operation of multiplying the parameter I of the vertex by d / z. Even if the parameter I of the vertex is finite, z
Values can be infinite in three-dimensional world coordinates. Therefore, the calculation result of dividing the parameter by the z value also requires infinite calculation accuracy, and it is difficult to estimate the calculation accuracy, which is a bad effect on hardware design. For example, world coordinates may have 2 ³⁸ accuracy of the calculation result of dividing the z value is 2 ^-38 to 2 ⁺³⁸ of accuracy is required. This high-precision operation requires large-scale hardware. It has been found that lowering this precision causes disturbance of the image displayed on the screen.

Accordingly, an object of the present invention is to provide an image processing apparatus, an image processing method thereof, and a recording medium on which the image processing program is recorded, capable of efficiently performing a perspective interpolation operation.

Another object of the present invention is to provide an image processing apparatus, an image processing method thereof, and a recording medium on which the image processing program is recorded, which makes it possible to easily design hardware for performing a perspective interpolation operation. .

[0020]

According to the present invention, as in the prior art, a polygon parameter obtained by a perspective transformation using a perspective value is interpolated by an internal division ratio on screen coordinates to obtain a pixel parameter. Instead of obtaining the ratio, the internal ratio on the world coordinates is once determined from the perspective value and the internal ratio on the screen coordinates. Then, an interpolation calculation process is performed on the parameters of the polygon based on the internal division ratio on the world coordinate to obtain the parameters of the pixel. In such a perspective interpolation operation, a parameter of a pixel on a world coordinate can be directly interpolated.

As a result, it is possible to directly interpolate the parameters of the polygons on the world coordinates, thereby eliminating the conventional step of dividing by the perspective value in the perspective interpolation operation. Further, since the internal division ratio on the world coordinates can be calculated only in accordance with the accuracy in the screen screen, the arithmetic processing is light.

In order to achieve the above object, the present invention provides:
Image processing for perspectively transforming polygon data on world coordinates into polygon data on screen coordinates, obtaining the parameters of pixels to be displayed by performing interpolation calculation processing on the parameters of the polygons, and generating image data of the pixels from the parameters A perspective transformation unit that converts coordinate data on world coordinates of a plurality of points constituting the polygon into coordinate data on screen coordinates based on a perspective value indicating perspective on the screen coordinates; For the pixels within, determine the internal division ratio on the world coordinates from the internal division ratio on the screen coordinates from the plurality of points and the perspective value, and determine the parameters of the plurality of points based on the internal division ratio on the world coordinates. The parameters of the pixel are obtained by performing an interpolation operation, and the pixel And a rendering unit for generating image data.

Further, the above object is achieved by providing a recording medium on which an image processing method and an image processing program of the present invention are recorded.

[0024]

Embodiments of the present invention will be described below with reference to the drawings. However, the technical scope of the present invention is not limited to the embodiment.

First, the principle of the perspective interpolation calculation in the embodiment will be described. FIG. 1 is a diagram illustrating perspective interpolation according to the embodiment. In this embodiment, the internal division ratio t ′ in the world coordinates is obtained from the internal division ratio t and the z value in the screen coordinates, instead of directly performing the perspective interpolation calculation on the parameters given for each vertex. , And vertex parameters are directly subjected to linear interpolation using the internal division ratio t ′.

FIG. 1A shows a triangular polygon ABC in world coordinates. Each vertex is given a coordinate value (x, y, z) on world coordinates and a parameter I for generating image data. Pixel PX (sx,
In order to obtain the parameter Ip in sy), a raster scan method is used as in the conventional example. However, in the present embodiment, the vertex parameter I is not multiplied by the perspective value d / z to obtain the vertex parameter on the screen coordinates in consideration of the perspective as in the related art.

[0027] The parameters I _l and I _r Considering perspective of the left edge Ls and the right edge Rs shown in FIG. 1 (2) is calculated as follows. For example, as an example of the parameter I _l of the left edge Ls, the internal division ratio t on the screen coordinates
Is t = (sy−sy ₀ ) / (sy ₁ −sy ₀ ), and the left edge Ls is obtained by the perspective interpolation operation.
The parameter I _l of

[0028]

(Equation 1) This equation (1) is rearranged as follows.

[0029]

(Equation 2) Therefore,

[0030]

(Equation 3) I _l = I ₁ × t ′ + I ₀ × (1.0−t ′) (4) and an equation for directly linearly interpolating the parameters I ₀ and I ₁ using the internal division ratio t ′. Can be represented by

That is, as shown in FIG.
Parameter I in di Ls_lUses the internal division ratio t '
And the parameters of vertices A and B on world coordinates
I₀And I₁Can be directly subjected to linear interpolation. One
In other words, this internal division ratio t 'is called the internal division ratio on world coordinates.
can do. Parameter I at right edge Rs _r
Can be obtained in the same manner.
You. In addition, as is clear from the above equation (3), the world coordinates
The above internal division ratio t 'is the internal division ratio t on screen coordinates and the z value.
And can be obtained from:

The internal division ratio t 'on the world coordinates is first calculated and stored in the memory, and the subsequent perspective interpolation operation for a plurality of parameters I is performed using the stored internal division ratio t'. An interpolation operation is directly performed on the vertex parameter I as in equation (4). As a result, the operation of dividing by the last d / z in the conventional example becomes unnecessary. Further, although the calculation of the internal division ratio t ′ itself in consideration of the perspective involves division by an infinite range of z values, as shown in FIG. , And the range of the internal division ratio t 'is 0 to 1.0. Therefore, the scale of the hardware for the operation is not so large.

The parameters I _l of the edges Ls and Rs on both sides
When the I _r obtained as, as shown in FIG. 1 (4),
The parameter I _p at the pixel PX is determined similarly.
That is, the interior division ratio on the screen coordinates _{t = (sx-sx l)} / (sx r -sx l), determine the internal ratio t _'p on the world coordinate in the same manner as in the above formula (3). Its internal ratio t 'by using the _p, both edges point Ls, and interpolation operation parameters I _l and I _r in Rs, the parameter I _p at the pixel PX, I _p = I _r × t' _p + I _l × (1.0−t ′ _p ) (5) That is, it is as shown in FIG.

In this operation, the operation of dividing by the last perspective value d / z in the conventional example becomes unnecessary as in the above case, and high accuracy is required for the operation of the internal division ratio t ′ _p on the world coordinates. Not done.

FIG. 2 is a block diagram of the image processing apparatus according to this embodiment. The CPU 20 is connected to a RAM via a bus.
21, the ROM 22, and the interface 23. The ROM 22 stores a game program or a simulation program, and the CPU 20 calculates the program according to an operation input signal from the interface 23 to generate image data.

Such image data is, for example, polygon data and viewpoint data in world coordinates. The image data is supplied to the geometry conversion unit 24. The geometry conversion unit 24 performs perspective transformation on the coordinate data of each vertex, and calculates the coordinate data of the vertex on the screen coordinates,
Find the perspective value. The operation of this perspective transformation is, specifically,
This is an operation of multiplying the above-mentioned perspective value d / z. The geometry conversion unit 24 stores the perspective-transformed polygon data in the polygon buffer 25.

FIG. 3 is a diagram showing an example of such polygon data. FIG. 3 shows polygon ID0 and polygon ID1.
Is shown. Each polygon has data of screen coordinates (sx, sy) and perspective value d / z for each vertex. These position data are perspective-transformed data in consideration of perspective with the perspective value d / z. Further, in the example of FIG. 3, the parameters I used for generating the image data are texture coordinates (Tx, Ty), normal vectors (Nx, Ny, Nz), transparency α value, and color data. . In addition, it may include, for example, a luminance value. In the present embodiment, the perspective transformation processing for multiplying the parameter I by the perspective value is not performed.

The rendering processing unit 26 performs rendering processing for obtaining pixel image data using the polygon data stored in the polygon buffer 25. The rendering processing unit 26 includes an edge interpolator 27, a raster interpolator 29, a Z comparator 31, a Z buffer 32, an image data generator 33, and a texture map 34.

As shown in FIGS. 1 (2) and (3), the edge interpolator 27 calculates the perspective value (d at the edge points Ls and Rs on both sides.
/ Z) _l, (seek d / z) _r, the internal ratio of the screen coordinates t _l, t _r and its perspective value _{(d / z) l, (} d / z)
The internal division ratios t ′ _l and t ′ _r on the world coordinates are obtained from _r . The internal division ratios t ′ _l and t ′ _r are used as calculation parameters in the subsequent image processing. Then, using the internal division ratios t ′ _l , t ′ _r , a linear interpolation operation of the parameters I of the vertices on both sides is performed, and the parameters I _{l on} both edges are obtained.
And the I _r is required.

In the raster interpolator 29, FIGS.
Perspective value of the street, the pixel PX shown in (d / z) calculated for _p, on the screen coordinate internal ratio t _p and its perspective value (d
/ Z) _p and the internal division ratio t ′ _p on the world coordinates is obtained. This internal division ratio t ′ _p is used as a calculation parameter in the subsequent image processing. Then, using the internal division ratio t ′ _p , linear interpolation calculation of the parameters I _l and I _r at both edge points is performed, and the parameter I _p at the pixel PX is obtained.

The Z comparator 31 and the Z buffer 32 perform hidden surface processing between a plurality of overlapping pixels. That is, Z
The comparator 31 compares the d / z of the Z-buffer with an infinite initial value with a perspective value d / z indicating the depth of the pixel being processed. Write the value d / z to the Z buffer 32. Then, the parameter _{Ip of} the pixel is supplied to the image data generation unit 33.

The image data generator 33 reads out the texture data stored in the texture map 34 in accordance with, for example, the texture coordinates (Tx, Ty) obtained by the perspective interpolation operation, and reads the normal vectors (Nx, Ny). , Nz), the shading process for the light source is performed, and the blending process is performed using the transparent value α value and the color data. Then, image data to be displayed is generated and recorded at a corresponding pixel position in the frame buffer 35. The image is displayed on the display 36 according to the image data recorded in the frame buffer 35.
Will be displayed.

FIG. 4 is a flow chart of image processing using perspective interpolation. The steps in this flowchart are processed by the image processing apparatus in FIG. A detailed image processing step will be described with reference to FIG.

First, the CPU 20 generates image data on world coordinates (S10). Such image data is
It is generated by executing a game program according to an operation input. Such image data includes coordinate data and parameter data of the vertices of the polygon, and is supplied to the geometry conversion unit 24. The geometry conversion unit perspectively converts the coordinate data of each vertex to obtain vertex coordinate data sz, sy, d / z (or 1 / z) on the screen coordinates (S20). In this embodiment, there is no need to perform perspective transformation up to the vertex parameter I as in the conventional example.

The screen coordinates (sx, sy) of the vertices perspectively transformed by the geometry transformation unit 24 and the perspective value d / z
(Or 1 / z) and the parameter I are recorded in the polygon buffer 25. Referring to the data recorded in the polygon buffer 25, the rendering processing unit 26
Performs a rendering process for obtaining image data for each pixel.

Steps S 21 to S 23 in FIG. 4 are performed by the edge interpolator 27. First, as shown in FIG. 1 (2), the perspective values (d / z) _l and (d / z) _{r at} the edge points Ls and Rs on both sides are obtained by interpolation (S2).
1). This interpolation operation uses the internal division ratio t = (sy−sy ₀ ) / (sy ₁ −sy ₀ ) at the left edge point Ls, and (d / z) _l = (d / z) ₁ × t + (d / z) ₂ × (1
−t). Such a perspective interpolation operation is the same as the conventional example.

Using the perspective values (d / z) _l and (d / z) _r at the edge points Ls and Rs on both sides, FIG.
As shown in (3), the internal division ratios t ′ _l and t ′ _r on the world coordinates on both sides are obtained according to the above equation (3) (S2
2). Further, from the vertex parameter I, the internal division ratios t ′ _l and t ′
_{Using r} , the parameter I at both edge points Ls, Rs
and _l and I _r is obtained (S23). Such calculation is performed by the vertex parameters I a 'direct interpolation arithmetic expression using _{(4), I l = I} 1 × t' internal ratio t on the world coordinate _{+ I 0 × (1.0-t} ') .

Steps S24 to S26 in FIG. 4 are performed by the raster interpolator 29. First, as shown in FIG. 1 (4), the perspective value (d /
z) _p is determined (S24). The interpolation operation, the internal division ratio t = (sx-sx _l) in pixels PX / (sx _r -
sx _{l) using, (d / z) p =} (d / z) r × t + (d / z) l × (1
−t). Such a perspective interpolation operation is the same as the conventional example.

Then, the perspective value (d /
z) By using _p , as shown in FIG. 1 (5), an internal division ratio t ′ _p of the pixel is obtained according to the above equation (3) (S2).
5). Furthermore, by utilizing the internal ratio t _'p from the parameter I _l and I _r both edge points, the parameter I in pixels PX
_p is obtained (S26). This operation is performed by directly interpolating the parameters I _l and I _{r of} both edge points using the internal division ratio t ′ _p (5), I _p = I _r × t ′ _p + I _l × (1.0 −t ′ _p ).

As shown in FIG. 4, steps S24-S
Step 26 is repeated by x size between both edge points.
Steps S21 to S26 are the minimum sy of the polygon.
Is repeated for y sizes from to sy. The screen coordinates at each edge point by the above-described raster scan method are, for example, the difference accumulation method (Digital Differential of DDA).
Analyzer). That is, it is obtained by repeating the calculation of sx = sx ₀ + Δx sy = sy ₀ +1 on the screen coordinates (sx ₀ , sy ₀ ) of the vertex A using the side gradient Δx. By repeating such addition, it is possible to avoid multiplication that the hardware is not good at.

Finally, the image data generating unit 33 calculates image data from the obtained parameter _{Ip of} the pixel and writes the calculated image data into the frame buffer 35.

The image processing apparatus shown in FIG. 2 is provided with a Z comparator 31. The Z comparator 31 compares the perspective value (d / z) in the Z buffer 32 with the perspective value of the pixel being processed for hidden surface removal processing. The pixel parameter I is supplied to the image data generation unit 33. Therefore, step S1 in FIG.
5 performs an operation of generating the image data only for the pixel determined to be located on the near side in the screen by the Z comparator 31.

[Second Embodiment] The embodiment described above
In the configuration example, all pixels in the polygon are
Internal division ratio t 'on the field coordinates_l, T '_r, T '_pAsk for
With each parameter I_l, I_r, I_pAsk for
Was. Finally, a perspective value (d / z) is obtained by a Z comparator. _p
Was performed according to the following. Therefore, it is processed first
Pixel data is written to the frame buffer
After that, the perspective value (d /
z) is large (located in front), the pixel
Image data is stored in the frame buffer.
Written. Therefore, it is determined that drawing is unnecessary by the Z comparator.
Calculation of the internal division ratio t 'and parameters of the pixel to be
The calculation for generating the image data to be overwritten is useless.

In the second embodiment, no such calculation is performed.
To eliminate waste, all polygons in one frame
On the other hand, the internal division ratio t ′ on the world coordinate of each pixel_l,
t ' _r, T '_pJust ask first, those pixels far away
Near value (d / z)_pHidden surface removal processing by Z comparator based on
I do. Finally, it is displayed using the internal division ratio described above.
Pixel parameters from the vertex parameters
Calculate.

FIG. 5 is a block diagram of an image processing apparatus according to the second embodiment. The same parts as those in the block diagram of FIG. 2 are given the same reference numerals. As shown in FIG. 5, in the image processing apparatus according to the second embodiment, the edge interpolator 27 in the rendering processing unit 26 outputs the edge points L on both sides.
Perspective values (d / z) _l and (d / z) _r at s and Rs and their internal division ratios t ′ _l and t ′ _r are obtained. Further, in the raster interpolator 29, from the perspective values at the edge points on both sides,
Perspective value (d / z) _{p of} pixel and internal division ratio t 'in horizontal direction
_p is required. Edge interpolator 27 and raster interpolator 2
In No. 9, the interpolation calculation of the parameter I using the internal division ratio is not performed. Therefore, those calculation steps are simplified.

In the second embodiment, the Z comparator 31
The distance values being processed pixel (d / z) _p and perspective value of the corresponding pixel stored in the Z buffer (d /
z) is compared. The perspective value of the pixel being processed (d /
z) If _p is greater than the perspective value in the Z buffer, it means that the pixel being processed is located closer to the screen, and the perspective value (d / z) _p overwrites the Z buffer 32. Is done. At the same time, the internal division ratios t ′ _l , t ′ _r ,
t ′ _p and the polygon ID are similarly overwritten.

The above-described edge interpolator 27 and raster interpolator 2
9, writing to the Z comparator 31 and the Z buffer 32 is 1
When the processing is completed for all the polygons in the frame, the perspective value, the internal division ratio, and the polygon ID of the pixel of the polygon located at the forefront in the display screen are stored in the Z buffer 32.

Therefore, the interpolator 40 calculates the perspective value (d / z) _p , the internal division ratio t ′ _l ,
t ′ _r , t ′ _p and the polygon ID are read out for each pixel, and the parameter I _p of the pixel is obtained from the parameters of the vertices of the polygon recorded in the polygon buffer 25.
The calculation is performed by linear interpolation using the above internal division ratios t ′ _l , t ′ _r , and t ′ _p . That is, a linear interpolation operation is performed using the internal division ratio on the world coordinate, which is obtained and stored in advance using the parameters on the world coordinate.

Then, the parameter I _{p of} each pixel obtained by the interpolator 40 is used as the image data generation unit 33.
The image data generating unit 33 generates a texture map 34 according to a texture coordinate which is one of the parameters.
, The shading process is performed using the normal vector which is one of the parameters, and the blending process is performed using the transparency α value.
It goes without saying that color data is used for the above processing. The image data generated by the image data generation unit 33 is
The data is written to the frame buffer 35 and the display 36
Will be displayed.

As described above, in this embodiment, the hidden surface erasing process is first performed using the perspective value, and then the interpolation calculation of the parameters for each pixel and the calculation of the image data are performed. The parameter value and the image data are not wasted, and efficient rendering processing is enabled.

FIG. 6 is a flowchart of the image processing according to the second embodiment. The same processing steps as those in the flowchart of FIG. 4 have the same step numbers. As shown in FIG. 6, steps S10 and S20 are the same as those in FIG. That is, first, the CPU 20 generates image data on world coordinates (S10). Such image data is
It has position data and parameter data of the vertices of the polygon, and is supplied to the geometry conversion unit 24. The geometry conversion unit performs perspective conversion of each vertex parameter to obtain vertex coordinate data sz, sy, d / z on screen coordinates (S20). The vertex parameter I is not perspective transformed. Then, the screen coordinates (sx, sy) of the vertices subjected to perspective transformation by the geometry transformation unit 24 and the perspective value d /
z and the parameter I are recorded in the polygon buffer 25. With reference to the data recorded in the polygon buffer 25, the rendering processing unit 26 performs a drawing process for obtaining image data for each pixel.

In the rendering processing section 26, first, the edge interpolator 27 obtains the perspective values (d / z) _l and (d / z) _r at the two side edge points Ls and Rs (S21),
Further, the internal division ratio t 'at both edge points shown in the above equation (3)
_l and t ′ _r are obtained (S22). At this point,
No interpolation calculation of parameters at both edge points is performed.

Further, in the raster interpolator 29, the perspective value (d / z) _p at the pixel PX on the horizontal line between the two edge points is obtained.
(S24), and further, the internal division ratio t ′ _p at the pixel PX.
Is obtained (S25). At this point, no interpolation calculation of the pixel parameters is performed.

Then, the Z comparator 31 compares the perspective value (d / z) _p of the pixel being processed with the perspective value recorded in the Z buffer 32 (S30). When the perspective value of the pixel being processed is large (located on the front side of the screen), the perspective value (d / z) _{p of} the pixel, the internal division ratio t ′ _l obtained by the edge interpolator and the raster interpolator, t ′ _r , t ′ _p and their polygon IDs are written into the Z buffer 32 (S3
1).

The above steps S24, S25, S30, S31
Is repeated by x size between both edges. Steps S21, S22 and S24, S25, S30,
31 is repeated for the y size from the vertex having the minimum sy value to the vertex having the maximum sy value. Also, their processing
This is repeated for all polygons in one frame.

When the above processing is completed for all polygons in one frame, the interpolator 40 proceeds to step S32
34, t'l, t'r,
t'p and the polygon ID are read (S32), and the parameters of the vertices corresponding to the polygon ID, Tx, Ty, α, Nx, Ny, Nz, etc. are read from the polygon buffer 25 (S33),
Vertex parameters, Tx, Ty, α, Nx, Ny, Nz, etc.
An interpolation operation is performed using l, t'r, and t'p (S34).

Then, the image data generation unit 33 obtains the color of the pixel PX from the parameters Tx, Ty, α, Nx, Ny, Nz, etc. obtained by the interpolation calculation (S35). More specifically, the processing includes reading of texture data, shading processing, blending processing, and the like. The image data having the color data of the pixel PX is stored in the frame buffer memory 35.
(S36). Steps S32 to S36 above
Is repeated for pixels. Finally, the image data in the frame buffer memory 35 is supplied to the display device 36 and displayed.

As described above, in the second embodiment,
After the hidden surface removal processing of a plurality of polygons is completed, the parameters of the vertices are directly linearly interpolated according to the internal division ratios t ′ _l , t ′ _r , and t ′ _p on the world coordinates in which the parameters of each pixel are obtained in advance. Then, the parameters of the pixel are obtained, and the arithmetic processing of the image data having the color data is performed based on the parameters. Therefore, there is no useless calculation, and the efficiency of the rendering process is improved.

[Image Processing by General-Purpose Computer] FIG.
FIG. 3 is a diagram showing a configuration example in a case where image processing of the present embodiment is performed by software using a general-purpose computer. When image processing is performed using a general-purpose computer, image processing calculations are performed according to a program stored in a recording medium. Therefore, the general-purpose computer operates as a computer dedicated to image processing by storing the image processing program in a computer-readable recording medium. The image processing program causes a computer to execute the procedures described in the flowcharts and the like.

In the example of FIG. 7, a CPU 70, a RAM 71 for calculation provided in a RAM 78, and a memory 72 storing a game program and an image processing program
6 is connected. The input / output unit 73 connected to the bus 76 is connected to an operation unit 74 operated by an operator, and inputs an operation signal. For image processing, a polygon buffer 25 for storing polygon data and a Z buffer 32 are provided in the RAM 78, and are connected to the bus 76, respectively. Further, the frame buffer memory 35 is connected to the bus 76 and also to the external display device 36.

In this example, the image processing program is stored in the memory 72.
An image processing program can be installed in the RAM 71 from a recording medium 75 such as an OM or a magnetic tape. Alternatively, the image processing program may be supplied by a communication medium.

In the above embodiment, the perspective transformation is performed using the perspective value d / z. However, the present invention
Perspective transformation can also be performed by setting the distance from the viewpoint to the screen to be a constant 1 and using 1 / z as the perspective value. In that case, the angle of view is uniquely determined. Therefore, the same result as in the above embodiment can be obtained by uniformly multiplying the screen coordinates (sz, sy) by the d value according to the display magnification after performing the perspective interpolation. Note that the internal division ratio t ′ on the world coordinates can also be obtained from the internal division ratio t and 1 / z on the screen coordinates as shown in Expression (3).

[0073]

As described above, according to the present invention, when perspective transformation is performed to obtain a perspective, and when the vertex parameters of a polygon are interpolated while maintaining the perspective, the above equation ( By obtaining and using the internal division ratio t ′ on the world coordinates according to 3), it is possible to directly perform a linear interpolation operation on the vertex parameters. As a result, it is possible to avoid the division of the parameters of each pixel by the last perspective value in the perspective interpolation calculation as in the conventional example. Furthermore, since the internal division ratio t 'requires only the accuracy on the screen coordinates, the hardware scale of the arithmetic circuit can be reduced. If the processing is performed by software, the processing time can be shortened.

Further, according to the present invention, the internal division ratio of each pixel is obtained first, the hidden surface elimination processing of a plurality of polygons is performed based on the perspective value d / z or 1 / z, and finally, according to the internal division ratio. By performing the interpolation calculation of the parameter of the pixel to be displayed and the calculation of the image data, it is possible to omit useless calculation of the rendering process.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating perspective interpolation according to an embodiment;

FIG. 2 is a block diagram of the image processing apparatus according to the embodiment.

FIG. 3 is a diagram showing an example of polygon data.

FIG. 4 is a flowchart of image processing using perspective interpolation.

FIG. 5 is a block diagram of an image processing apparatus according to a second embodiment.

FIG. 6 is a flowchart of image processing according to the second embodiment.

FIG. 7 is a diagram illustrating a configuration example in a case where image processing of the present embodiment is performed by software using a general-purpose computer.

FIG. 8 is a diagram illustrating perspective projection.

FIG. 9 is a flowchart of image processing using conventional perspective interpolation.

FIG. 10 is a diagram illustrating perspective interpolation.

[Explanation of symbols]

 24 geometry conversion unit (perspective conversion unit) 26 rendering processing unit 27 edge interpolator 29 raster interpolator 31 Z comparator 32 Z buffer 35 frame buffer 36 display

Claims (9)

[Claims] 1. Polygon data on world coordinates is perspective-transformed into polygon data on screen coordinates, and parameters of the polygon are interpolated to obtain parameters of a pixel to be displayed. From the parameters, image data of the pixel is obtained. A perspective transformation for converting coordinate data on a world coordinate of a plurality of points constituting the polygon into coordinate data on a screen coordinate based on a perspective value indicating a perspective on the screen coordinate. Part, for the pixels in the polygon, determine the internal division ratio on the world coordinates from the internal division ratio on the screen coordinates from the multiple points and the perspective value, and set the parameters of the multiple points on the world coordinates. Interpolation processing is performed based on the internal division ratio to determine the parameters of the pixel, and the parameters of the pixel The image processing apparatus and a rendering processor for generating image data of Luo the pixel. 2. The world coordinate system according to claim 1, wherein the perspective value is d / z or 1 / z (z is a z value on world coordinates) and the internal division ratio on the screen coordinates is t. The above internal division ratio t ′ is given by t ′ = (d / z ₁ ) · t / ｛(d / z ₁ ) · t + (d /
_{z 0) · (1-t} )} or _{t '= (1 / z 1} ) · t / {(1 / z 1) · t + (1 /
z ₀ ) · (1−t)｝ (where z ₀ and z ₁ are z values on both sides of the pixel). 3. The polygon processing device according to claim 1, wherein the plurality of points are vertices of the polygon, and the rendering processing unit obtains an internal division ratio on the world coordinate for an edge point on a side between the vertices. The parameters of the vertices are interpolated to obtain the parameters of the edge points based on the internal division ratios. Further, the internal division ratios on the world coordinates of the pixels between the edge points are obtained. An image processing apparatus for performing an interpolation calculation process on the parameters of the edge point based on the parameter to obtain a parameter of the pixel. 4. The image processing apparatus according to claim 1, wherein the internal division ratio on the world coordinate is obtained with an accuracy according to an accuracy in a display screen. 5. The rendering processing unit according to claim 1, wherein the rendering processing unit performs a hidden surface removal process on a plurality of polygons according to a perspective value of the pixel, and performs an interpolation operation process based on the internal division ratio after the hidden surface removal process. An image processing apparatus, wherein a parameter of a pixel is obtained by the following formula, and image data of the pixel is generated from the parameter of the pixel. 6. A perspective transformation of polygon data on world coordinates to polygon data on screen coordinates, interpolation calculation processing of the polygon parameters to obtain parameters of pixels to be displayed, and image data of the pixels from the parameters. A perspective transformation for converting coordinate data on world coordinates of a plurality of points constituting the polygon into coordinate data on screen coordinates based on a perspective value indicating perspective on the screen coordinates. Determining the internal division ratio on the world coordinates from the internal division ratio on the screen coordinates from the plurality of points and the perspective value for the pixels in the polygon, and changing the parameters of the plural points on the world coordinates. Interpolation processing is performed based on the internal division ratio to obtain the parameters of the pixel, and the parameters of the pixel are calculated. Image processing method and a rendering step of generating image data of the pixels from. 7. The world coordinate according to claim 6, wherein the perspective value is d / z or 1 / z (z is a z value on world coordinates) and the internal division ratio on the screen coordinates is t. The above internal division ratio t ′ is given by t ′ = (d / z ₁ ) · t / ｛(d / z ₁ ) · t + (d /
_{z 0) · (1-t} )} or _{t '= (1 / z 1} ) · t / {(1 / z 1) · t + (1 /
z ₀ ) · (1−t)｝ (where z ₀ and z ₁ are z values on both sides of the pixel). 8. Perspective transformation of polygon data on world coordinates to polygon data on screen coordinates, interpolation processing of the polygon parameters to obtain parameters of pixels to be displayed, and image data of the pixels from the parameters. In a recording medium on which an image processing program for generating the polygon is generated, coordinate data on world coordinates of a plurality of points constituting the polygon is converted into coordinate data on screen coordinates based on a perspective value indicating perspective on the screen coordinates. And a perspective transformation procedure for transforming, into the pixels in the polygon, an internal division ratio on world coordinates from the internal division ratio on the screen coordinates from the plural points and the perspective value, and the parameters of the plural points are obtained. Interpolation processing is performed based on the internal division ratio on the world coordinates to obtain the parameters of the pixel. The computer readable recording medium recording an image processing program for processing and rendering processing procedure for generating image data of the pixels from parameters of the pixels on the computer. 9. The system according to claim 8, wherein the perspective value is d / z or 1 / z (z is a z value on world coordinates), and the internal division ratio on the screen coordinates is t. The above internal division ratio t ′ is given by t ′ = (d / z ₁ ) · t / ｛(d / z ₁ ) · t + (d /
_{z 0) · (1-t} )} or _{t '= (1 / z 1} ) · t / {(1 / z 1) · t + (1 /
z ₀ ) · (1-t)｝ (where z ₀ and z ₁ are z values on both sides of the pixel). The computer-readable recording medium recording an image processing program.