JPH0887606A - Stereoscopic image processing device - Google Patents

Abstract

(57) [Summary] [Object] An object of the present invention is to provide a stereoscopic image processing apparatus capable of performing simple perspective texture mapping by reducing the amount of calculation and the amount of hardware. [Structure] Triangular polygon end point information, screen memory 5 for storing Z value and polygon parameter memory address, polygon end point visual field coordinate value, mapping polygon projection direction, mapping pattern memory address value, and two sides A polygon parameter memory 40 for storing the reciprocal of the outer product value of the vector; and determining the projection direction of the polygon from the direction of the normal vector of the polygon, and the polygon contour processing device 20 and the internal processing device 30 only use two-dimensional elements. Then, the two-dimensional mapping pattern memory address is calculated by the inner product mapping method, the mapping pattern memory 7 is accessed, the image information of each pixel is read and displayed on the CRT 9.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a stereoscopic image processing apparatus for projecting and displaying a three-dimensional polyhedron object on a two-dimensional screen, and in particular, it provides pixel information of a desired area of a texture plane to a desired area of a display plane. The present invention relates to a mapping processing device that projects on a graphic data of a region.

[0002]

2. Description of the Related Art Two-dimensional (flat) of a CRT display or the like by subjecting a three-dimensional solid figure to perspective conversion processing, perspective processing, etc.
In the case of compositing and displaying on the screen, it is necessary for an object existing in the front to perform a process of hiding a part or all of the objects located behind it, that is, a hidden surface erasing process. As a method of hidden surface removal processing, Z sort method (painting over method), Z
A buffer method, a scan line method, etc. are known.

The Z-sort method has an advantage that the processing can be performed at a very high speed, but has a drawback that a model cannot be drawn correctly when polygons intersect each other.

Pixels within the processing of polygons
The Z-buffer method is an algorithm that makes a determination before and after in units of 10 to eliminate the drawbacks of the Z-sort method.
In the Z buffer method, the color data of the polygon to be displayed in each pixel and the depth of the surface, that is, the Z value (distance from the starting point) of the object is stored in pixel units, and is stored each time a new polygon is input. The Z value of the new polygon is compared with the current Z value, and the Z value is updated only when the new Z value is smaller.
At the same time, new polygon color data is registered. This allows distant objects to be overwritten by nearby objects, resulting in an image with hidden surfaces removed.

This Z-buffer method requires a Z-buffer memory for storing the Z-value of an object for each pixel, and has a problem that a large image memory having a size corresponding to the number of pixels is required as a whole.

On the other hand, in the scan line method, when color data of each pixel is displayed for each raster scan like CRT, the adjacent pixel, that is, the pixel of the next scan line, has a very strong correlation with the current pixel. This method is suitable for a device that performs sequential processing by paying attention to the fact that it has, but has a drawback that it requires many calculations and the control logic becomes complicated.

As an intermediate hidden surface processing of the above two hidden surface removal processing methods, the correlation between scan lines is used to
An apparatus using the Z buffer method in the line is disclosed in, for example, Japanese Patent Laid-Open No. 62-100878. This includes "a depth register that holds the depth distance (Z value), a brightness register that holds the brightness (color / luminance) data, inside / outside determination within the range of the plane segment, displacement addition of the depth distance, and depth data One adder for performing divisional comparison, an input / output means for outputting the information about the input plane segment token through a one-stage pipeline register, and a luminance data bus for outputting the contents of the luminance register to the outside. Hidden line processing device provided with ". According to this device, hidden surface processing can be performed with a small amount of hardware.

Further, texture mapping is performed in order to express patterns and irregularities on the surface of the object. As this mapping processing method, an algorithm for generating an image in real time has been proposed. For example, the Technical Report of the Television Society of Japan, TEBS105-9PPOE62-9 (198).
Details on 5) "3D special effect device".

In the above mapping processing method, first, some points in the input screen are mapped to points on the output screen by a shape generation operation, and some intermediate points on the output screen are calculated. By interpolating / inverse mapping calculation (interpolation mapping calculation) from, the points on the input screen are associated with each other. That is, the load of the shape generation calculation is reduced by performing the spatial interpolation using the continuity of the shapes.

Further, since the entire shape is calculated from the calculation result of a part of the shape, the load of the shape generation calculation can be further reduced by performing time interpolation on a small amount of intermediate data.

Further, there is an advantage that the filter control data and the interpolation control data can be easily calculated. The spatial interpolation / temporal interpolation calculation may be the same even if the mapping shape is changed.

The above spatial interpolation will be described. Utilizing that the shape surface is continuous at least locally,
A three-dimensional shape is approximated to a polygon (polyhedron), and the shape generation operation is performed only for the vertices.

As shown in FIG. 31, the mapping operation from the point on the output screen to the point on the input screen is performed by calculating the weighting factors K, L and M of the three vertices from the vertex coordinates of the triangular plate-shaped polygon. A corresponding point on the input screen is obtained by applying a weighting factor to the vertex coordinates of the input screen.

That is, as shown in FIG. 32, each vertex is (AX, AY, AZ), (BX, BY, BZ), (C
X, CY, CZ), (PX, PY, PZ), the corresponding points on the input screen are calculated based on the flowchart shown in FIG.

In this calculation, first, the distances between the vertices are calculated (step S100), and the vectors AB and AC are calculated.
The outer product of is calculated (step S101). Then, the distance to the corresponding point (P) is calculated (step S102), and the cross product of the vectors between the corresponding point and each vertex is obtained. That is, the outer product of the vector PB and the vector PC (step S103), the outer product of the vector PC and the vector PA (step S104), and the outer product of the vector PA and the vector PB (step S105) are obtained. Next, the weighting factors K, L, and M of the three vertices are calculated (step S106), and the weighting factors are applied to obtain corresponding points (step S10).
7).

As the front / back determination flag for erasing the hidden surface, the sign of the z component of the product of the AB vector and the AC vector is used to determine whether the surface is facing the viewpoint side or the opposite side. The mapping operation is performed only on the surface facing the viewpoint.

[0017]

However, the above-mentioned interpolation inverse mapping operation, that is, the interpolation mapping algorithm, uses the formula shown in FIG. 31, which is used to perform perspective perspective mapping in the reverse direction. When the two-dimensional mapping pattern coordinates are obtained from the polygon coordinates, the calculation is based on the flowchart shown in FIG. 33, which requires a large number of calculations, and therefore requires a lot of hardware for real-time mapping. was there.

According to the present invention, by performing the interpolation mapping processing for obtaining the two-dimensional mapping pattern coordinates from only the two-dimensional elements, the calculation amount is reduced and the square root calculation which requires a large amount of hardware is applied. An object of the present invention is to provide a stereoscopic image processing device capable of processing simple perspective texture mapping in real time at low cost.

[0019]

According to the present invention, there is provided storage means for storing end point information of a mapping pattern memory for storing end point information constituting a triangular polygon and a texture image and depth information of the polygon, and the storage means. The coordinate information of each end point information is converted into the screen end point information, and the conversion means for selecting the projection direction of the mapping polygon according to the direction of the normal vector on the polygon field of view coordinates, and the polygon polygon based on the projection direction of the mapping polygon. A means for calculating the reciprocal of the cross product value of the vectors of the two sides, the field coordinate value of the end point of the polygon, the end point information of the mapping pattern memory, the projection direction of the mapping polygon, and the reciprocal number of the cross product value of the vectors of the two sides are stored. The polygon parameter memory, screen end point information from the above conversion means, and the depth of the polygon Screen memory for storing the address information of the polygon parameter memory and the address information of the polygon outline based on the screen end point information from the conversion means, and the address value of the polygon parameter memory and the address value of the polygon parameter memory for each scan line. A contour processing means for converting the contour portion information of the polygon, a means for computing the information between the two opposite sides calculated by the contour processing means, and calculating the displacement of the displacement depth information of the X address, and the scan. A means for determining whether or not the pixel position corresponding to each pixel of the line exists within the range of the polygon, and the depth information of the pixel at the forefront and the depth information of the polygon at the pixel position are compared and compared. To rewrite the depth information of the polygon to the depth information of the polygon that always exists in the foreground , Means for adding the displacement of the depth information to the depth information of the polygon to calculate the depth information of the adjacent pixel position, and at least one X between the two opposite sides of the polygon existing at the front at the compared pixel position. A means for calculating an address and an address value of the polygon parameter memory, and a weighting coefficient of three vertices of the polygon based on the information of the polygon parameter memory read by accessing the calculated address value of the polygon parameter memory, It is provided with means for calculating an address value of the mapping pattern memory, and means for accessing the mapping pattern memory based on the calculated mapping pattern address and transferring the image data to the display means.

Further, the projection direction of the mapping polygon may be selected according to the size of the scalar quantity of the normal vector on the polygon visual field coordinates.

[0021]

According to the present invention, the projection direction of the mapping is X- from the direction of the normal vector of the three-dimensional polygon on the visual field coordinates.
It is determined whether it is Z, YZ, or XY. One of them is selected, and interpolation mapping processing for obtaining the two-dimensional mapping pattern coordinates from only the two-dimensional element is performed. As a result, by performing a square root calculation that requires a large amount of hardware, the calculation process can be largely omitted, and simple perspective texture mapping can be processed in real time at low cost.

[0022]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the overall configuration of a pseudo three-dimensional image processing apparatus using the present invention. This apparatus is suitable for use in a game machine such as a racing game or an airplane control simulation. An example is shown. The overall configuration of the present invention will be described with reference to FIG.

The structure of the image information supply device 10 will be described. This device comprises a world memory 1, a geometric conversion device 2, an operation section 3, and a CPU 4. In this world memory 1, all objects are represented as a collection of a plurality of polygons, and the end points of the polygons are represented by X, Y, and
It is stored as the Z coordinate. Further, in this world memory 1, the X of the polygon end point on the coordinates of the object,
Corresponding to the Y and Z coordinates and the polygon, the end point information data of the mapping pattern memory that stores the texture image is stored. The operation unit 3 includes a steering wheel, an accelerator, a brake, and the like, and the scanning content thereof is converted into an electric signal and output to the CPU 4.

The CPU 4 calculates the situation data according to this situation in accordance with the electric signal converted based on the contents of the operation unit 3 constituted by handle access or the like, and gives the data to the geometrical transformation device 2.

The geometrical transformation device 2 reads the data of the end point information of each polygon from the world memory 1 in accordance with the instruction from the CPU 4, executes the matrix calculation necessary for the movement of the object and the rotation of the visual field, and screens the end points of the world coordinates. Geometric transformation such as projection transformation is performed on the coordinates, and the two-dimensional X, Y screen data is given to the screen memory 5. In addition, the representative value of the polygon subjected to the field-of-view conversion, that is, the distance from the starting point of the polygon, that is, the depth distance data (Z value) is determined, and the data is given to the screen memory 5. As shown in FIG. 25, the screen memory 5 stores X and Y screen coordinate values and Z values for the end points of each polygon.

Further, the geometric transformation device 2 uses the X, Y, Z values on the visual field coordinates of the polygon end points, the X, Y coordinate (MX, MY) values of the mapping pattern memory, and the three-dimensional polygon on the visual field coordinates. From the direction of the normal vector or the like, it is selected which of the three types, XZ, YZ, and XY, the projection direction of the mapping polygon indicates, and a flag indicating the type is given to the polygon parameter memory 40.

This geometric transformation device 2 calculates a polygon normal vector, pays attention to the direction in which the X, Y, Z scalar amounts of the normal vector are the largest, and if it is large in the Y direction,
If it is large in the X-Z and X-directions, it is the Y-Z, and if it is large in the Z-direction, it is the direction in which the XY is projected, and a flag indicating that information is given to the polygon parameter memory 40. This polygon parameter memory 40, as shown in FIG.
For each end point of each polygon, the X, Y, and Z visual field coordinate values (E
X, EY, EZ), X, Y of the mapping pattern memory
Coordinate (MX, MY) value, mapping direction flag and 2
The reciprocal 1 / ab of the vector cross product of the sides is stored.

The polygons of the present invention are triangular polygons, and the end points of the polygons are, as shown in FIG.
X, AY, AZ), (BX, BY, BZ), (CX, C
Y, CZ).

That is, when the mapping direction is XY, the reciprocal of the vector cross product of the two sides 1 /
Calculate ab.

[0031]

1 / ab = (AX−BX) × (AY−CY) −
(AY-BY) x (AX-CX)

When the mapping direction is XZ, the reciprocal 1 / ab of the vector cross product of the two sides is calculated based on the following mathematical formula 2.

[0033]

2 / ab = (AX−BZ) × (AX−CX) −
(AX-BX) x (AZ-CZ)

When the mapping direction is YZ, the reciprocal 1 / ab of the vector cross product of the two sides is calculated based on the following equation 3.

[0035]

## EQU3 ## 1 / ab = (AY-BY) × (AZ-CZ)-
(AZ-BZ) x (AY-CY)

1 / ab calculated based on each of the above formulas
Are stored in the polygon parameter memory 40.

The polygon outer shape processing device 20 comprises a polygon extraction device 21, a parameter calculation device 22, and a vertical interpolation calculation device 23. Based on the XY addresses of the polygon end points read from the screen memory 5, the polygon extracting device 21 determines to which direction the vector of each side forming the polygon belongs, as shown in FIG. Accordingly, as shown in FIG. 29, it is determined whether the endpoints forming the sides of the polygon belong to the right side or the left side. Then, in the polygon extracting device 21, the end points of the respective sides forming the polygon from the screen memory 5, that is, X
The start point address (XS), end point address (XE), Y start point address (YS), right side address (YE), polygon depth distance data (Z value), and each data is stored in the parameter calculator 22. give.

The parameter arithmetic unit 22 of the parameters of the polygon outline shape processing unit 20 calculates the parameters required for obtaining the outline end point information of the polygon by digital differential analysis (DDA), and the parameters are calculated by the vertical interpolation arithmetic unit. Give to 23. The vertical interpolation calculation device 23 calculates by interpolating the outer edge point information and the Z value of each of the left side and the right side where the polygon intersects each scan line. Each calculated data is given to the polygon edge memory 6. Details of the polygon outer shape processing device 20 will be described later.

Then, in the polygon edge memory 6, as shown in FIG. 27, each data given from the polygon outer shape processing device 20, that is, the left side, right side value and left side Z value of the polygon for each scan line, The Z value on the right side is stored for each resolution in the vertical direction (Y address direction) of the screen.

Further, the polygon edge memory 6 has
Number of polygons stored in one Y address (CN
T) is written. That is, 1 for each Y address
Each time the number of polygons is stored, the number of polygons is counted up, and this count number (CNT) is written in the polygon edge memory 6.

Each data stored in the polygon edge memory 6 is given to the polygon internal processing device 30. The internal polygon processing device 30 includes a parameter calculation device 31,
It is composed of a horizontal interpolation calculation device 33 and a hidden surface processing device 34.

The parameter calculation unit 31 receives the values (XL, XR) of the left side X and the right side X of the polygon and the Z values (ZL, ZR) of the left side and the right side of the polygon from the polygon edge memory 6 for each scan line, and horizontally. The parameters required for the interpolation calculation are calculated, and the parameters are transferred to the hidden surface processing device 34.

The hidden surface processing unit 34 in this embodiment is Z
The hidden surface processing using the buffer method is performed. The parameters are received from the parameter calculation device 31 for each scan line, and the hidden surface processing is sequentially performed for each dot (pixel) on the pipeline. When all the data of one scan line are processed, the polygon parameter memory address of each dot is sequentially calculated by the horizontal interpolation calculation device 33.
The hidden surface is processed by transferring to. That is, at each dot, the Z value of the stored polygon in the foreground is compared with the Z value of the polygon to be processed, and if the Z value of the polygon to be processed is small, the parameter address of that polygon is set horizontally. The Z value is transferred to the interpolation calculation device 33, and the Z value is stored as the Z value to be compared. When the Z value of the polygon to be processed is large, the Z value and the parameter address are not rewritten, and the previously stored polygon parameter address value is used as the horizontal interpolation calculation device 3
3 is transferred.

The horizontal interpolation calculation device 33 is a hidden surface processing device 3.
The polygon parameter memory 40 is accessed according to the polygon parameter address received from No. 4, the parameter of each polygon is read, and the interpolation mapping operation is performed based on this parameter to calculate the mapping pattern memory address (MX, MY). At the calculated mapping pattern memory address (MX, MY), for example, by accessing the mapping pattern memory 7 in which the data shown in FIG. 30 is stored, each data is stored in the mapping pattern memory 7 based on the data stored in the mapping pattern memory 7. R, G, B or brightness (LUT) values of dots are read out sequentially,
Write to the frame memory 8. Details of the internal polygon processing device 30 will be described later.

The R, G, B or LUT values of the dots given from the polygon internal processing unit 30 to the frame memory 8 are transferred to the CRT 9 and displayed as an image.

Next, the polygon contour processing apparatus 2 of the present invention
0, the internal polygon processing device 30 is shown in FIGS.
Will be described with reference to.

In this embodiment, the polygon is a screen end point coordinate (X, Y) and a visual field coordinate (EX, EY, E).
Z) and the texture, that is, by having the end point coordinates (MX, MY) of the mapping pattern and the depth distance (Z value), the hidden surface processing by the Z buffer method is performed and, for example, the texture shown in FIG. It is transformed corresponding to the polygon surface and is mapped on the polygon surface.

First, the polygon outline processing performed by the polygon outline processing device 20 will be described with reference to FIGS.

For this outer shape processing, in the polygon extracting device 21, the vector of each side forming the polygon is determined in which direction shown in FIG. 28 based on the XY address of the polygon end point read from the screen memory 5. 29, it is determined whether the endpoints forming the sides of the polygon belong to the right side or the left side, as shown in FIG. 29, depending on the direction of the vector.

The screen memory 5 stores screen end point coordinates (X, Y) and the Z value of the polygon. Polygon extraction device 21 of polygon outline processing device 20
Accesses the screen memory 5, reads the screen end point coordinates (X, Y) and the Z value of each polygon from the screen memory 5, and calculates the outer shape of the polygon by digital differential analysis (DDA). The data corresponding to each side is given to the parameter calculation device 22 which determines the left side and calculates the parameter. That is, the X parameter calculation device 22a has screen coordinates, and the Z parameter calculation device 22b has screen coordinates (X, Y) and Z values (ZS, ZE) at the start and end points of each side.
Is given.

In the X parameter calculation device 22a of the polygon outline processing device 20, the starting point (YS) of the Y address of each side of the screen coordinates given by the polygon extraction device 21.
And a distance (DY) in the Y direction from the end point address (YE). That is, the calculation of DY = YE-YS is performed. Subsequently, DX = X in order to calculate the distance (DX) from the X end point (XE) to the X start point (XS) of each side of the screen coordinates.
E-XS calculation is performed. Using these DY and DX, DX / DY is calculated as a parameter for digital differential analysis (DDA) of the outer shape of the polygon, and the fine difference value (DD
X) is calculated. This DDX is transferred to the X-value vertical interpolation device 23a. Then, the X-value vertical interpolation device 23a performs interpolation calculation, and the interpolated data is stored in the polygon edge memory 6 via the polygon edge memory controller 24.

That is, as shown in equation (1) below,
The X-value parameter calculation device 22a calculates the fine difference value, and as shown in the following equation (2), the X-value vertical interpolation calculation device 23a performs interpolation calculation to determine the address from the start point to the end point of each side. calculate. The initial value of X in the equation (2) is the starting point data (XS).

[0053]

(4) DDX = (XE−XS) / DY (1) X = X + DDX (2)

The distance (DY) in the Y direction is calculated from the start point (YS) and end point address (YE) of the Y address of each side of the screen coordinates given by the polygon extraction device 21. That is, the calculation of DY = YE-YS is performed.

Furthermore, the Z value start point address (ZS) and end point address (ZE) read from the screen memory 5
The data for the polygon is calculated from the address data of the above by the digital differential analysis (DDA) in the Z value parameter computing device 22b and the Z value vertical interpolation device 23b based on the following equations (3) and (4). Stored in the edge memory 6. That is, the fine difference value is calculated as shown in Expression (3), and the interpolation calculation is performed as shown in Expression (4) to calculate the data from the end point to the start point of each side. The initial value in the equation (4) is the starting point data (XS).
The calculation of equation (4) is repeated from 0 to DY.

[0056]

## EQU5 ## DDZ = (ZE-ZS) / DY (3) Z = Z + DDZ (4)

In this embodiment, In synchronization with the scan line, the outline of the polygon, the outline address information of the mapping pattern modified based on the outline of the polygon, and the outline address information of the Z value are stored in the polygon edge memory 6 for each Y address.

Each of the above devices is controlled by a controller 25, and this controller 25 controls the parameter calculation device 22 and the vertical interpolation calculation device 23 in order to interpolate between the end points by DDA according to the flow chart shown in FIG.

Next, a concrete configuration example of the polygon outer shape processing device 20 is shown in FIGS. FIG. 3 is a circuit diagram showing a specific configuration of the X parameter calculation device 22a.

The X value parameter computing device 22a shown in FIG.
Is for calculating an X value parameter used for vertical interpolation calculation.

The start point (YS) read from the screen memory 5 is stored in the register 201, and the Y end point (YE) read from the screen memory 5 is stored in the register 202. From these registers 201 and 202, YS and YE are stored. Is input to the subtractor 205.

The subtractor 205 subtracts YS from YE, and this value (DY) is temporarily stored in the register 207.

The X start point (XS) read from the screen memory 5 is stored in the register 203, and the X end point (XE) read from the screen memory 5 is stored in the register 204.
And XS and XE are applied to the subtracter 206 from the registers 203 and 204.

The subtractor 206 subtracts XS from XE, and the register 208 temporarily stores this value (DX).

From the registers 207 and 208 to the divider 209
In addition, the DX processed by the subtractor 206 and the subtractor 205
The subtracted DY is given, and the value of DX is divided by DY. Fine difference value DD calculated by the divider 209
The X is temporarily stored in the register 210 and then transferred to the X vertical interpolator 23a.

FIG. 4 is a circuit diagram showing a specific structure of the Z parameter calculation device 22b. The Z parameter calculation device 22b shown in FIG. 4 is for calculating Z parameters used for vertical interpolation calculation.

The Y start point (YS) read from the screen memory 5 is stored in the register 235, and the Y end point (YE) read from the screen memory 5 is stored in the register 236. From these registers 235, 236, YS, YE is input to the subtractor 239.

The subtractor 239 subtracts YS from YE, and this value (DY) is temporarily stored in the register 241.

The Z start point (ZS) read from the screen memory 5 is stored in the register 237, and the Z end point (ZE) read from the screen memory 5 is stored in the register 238.
And XS and ZE are applied to the subtractor 240 from both the registers 237 and 238.

The subtractor 240 subtracts ZS from ZE, and this value (DZ) is temporarily stored in the register 242.

From the registers 241, 242 to the divider 243
Is given to DZ subjected to the subtraction processing by the subtractor and DY subjected to the subtraction processing at the subtractor 239, and the value of DZ is divided by DY. The fine difference value DDZ calculated by the divider 243 is temporarily stored in the register 244 and then transferred to the Z vertical interpolation device 23b.

Next, the configuration of the X-value vertical interpolation calculation device 23a will be described with reference to FIG.

The screen coordinate XS transferred from the X value parameter calculation device 22a is stored in the register 247 via the multiplexer 245.

The multiplexer 245 transfers the value of the X parameter calculation unit 22a to the register 247 only when receiving the start signal, and otherwise, the adder 24
8 is controlled to be transferred to the register 247.

The register 246 temporarily stores the value of the parameter DDX transferred from the X value parameter arithmetic unit 22a.

Upon receiving the start signal, the value of the register 246 is transferred to the adder 248. X and DDX are added by the adder 248, and the addition result (X) is given to the register 249 and stored in the polygon edge memory 6 via the polygon edge memory controller 24.

The structure of the Z-value vertical interpolation device 23b will be described with reference to FIG.

The ZS transferred from the Z value parameter computing unit 22b is transferred to the register 26 via the multiplexer 260.
Stored in 2.

The multiplexer 260 transfers the value of the Z value parameter arithmetic unit 22b to the register 262 only when receiving the start signal, and otherwise, the adder 26
3 output is controlled to be transferred to the register 262.

The register 261 temporarily stores the value of the parameter DDZ transferred from the Z value parameter calculation device 22b.

Upon receiving the start signal, the value of the register 262 is transferred to the adder 263. The adder 263 adds Z and DDZ, and the addition result (Z value) is given to the register 264 and stored in the polygon edge memory 6 via the polygon edge memory controller 24.

FIG. 7 shows the internal polygon processing device 30.
19 to FIG. Internal polygon processing unit 3
As described above, 0 is composed of the parameter calculation device 31, the horizontal interpolation calculation device 33, and the hidden surface processing device 34. First, the parameter calculation device 33 will be described with reference to FIGS.

The parameter calculator 31 reads the XY address between the two sides corresponding to each scan line, that is, the left side and the right side from the polygon edge memory 6, and based on the read address information, (5) to
The address of each pixel dot inside the polygon is calculated as the hidden surface parameter address according to the equation (6).

In other words, in this embodiment, the polygon edge memory controller 31a synchronizes with the scan line scanning signal, and the X left side (XL) of two points indicating the outer shape of the polygon corresponding to the Y address as its vertical position.
And X right side (XR), Z value (ZL, ZR) and polygon parameter memory address are read from the polygon edge memory 6.

From the left and right side addresses of the X address read from the polygon edge memory 6, the distance (DXY) in the X direction is calculated as shown in equation (5).

Using this DXY, the hidden surface processing parameter computing unit 31b calculates the Z value read from the polygon edge memory 6 based on the equation (6) and used in the digital differential analysis (DDA).

[0087]

(6) DXY = XR-XL (5) DZ = (ZR-ZL) (6)

Then, as shown in FIG. 7, the polygon edge memory controller 31a causes the polygon edge memory 6 to scan line addresses (XL, XR) and Z values (Z) for the left and right sides of each Y address (scan line).
L, ZR) and read the hidden surface processing parameter calculation device 3
Transfer data to 1b respectively.

FIG. 8 shows the hidden surface processing parameter calculation device 3 described above.
It is a circuit diagram which shows the specific structure of 1b.

Polygon edge memory controller 31a
As a result, the scan line address (XL) on the left side of each Y address (scan line) is given to the register 320, and the scan line address (XR) on the right side is given to the register 321, and from these registers 320, 321 XL, XR
Is input to the subtractor 324.

The subtractor 324 subtracts XL from XR and temporarily stores this value (DXY) in the register 326.

Further, the left side (ZL) of the Z value is registered in the registers 322 and Z by the polygon edge memory controller 31a.
The right-hand side (ZR) of the value is given to the register 323, respectively, and the subtracter 325 is used from these registers 322 and 323.
Are given ZL and ZR.

The subtractor 325 subtracts ZL from ZR, and this value (DZ) is temporarily stored in the register 327.

The subtractor 325 subtracts ZL from ZR, and this value (DZ) is temporarily stored in the register 327.

From the registers 326 and 327 to the divider 324
The subtraction processed DXY and the subtractor 325 subtracted DZ are given, and the value of DZ is divided by DXY. The fine difference value DDZ calculated by the divider 328 is stored in the register 3
After being temporarily stored in 29, it is transferred to the hidden surface processing device 34 via the hidden surface processing device interface (I / F) 31c. The above circuits are controlled by the controller 40,
It operates according to the flowchart of FIG.

FIG. 9 shows a block diagram of the hidden surface processing apparatus.
The hidden surface processing device 34 performs hidden surface processing using the Z buffer method, and the parameter calculation device 3 is provided for each scan line.
The parameter memory address of each dot is sequentially horizontal when all the data of one scan line are processed by receiving the parameter from 1 and performing the hidden surface processing for each dot (pixel) in sequence on the pipeline. The hidden surface processing is performed by transferring to the interpolation calculation device 33. That is, at each dot, the Z value of the stored polygon in the forefront is compared with the Z value of the polygon to be processed, and if the Z value of the polygon to be processed is small, the polygon parameter memory address of that polygon is set. Is transferred to the horizontal interpolation calculation device 33 and the Z value is stored as the Z value to be compared. When the Z value of the polygon to be processed is large, the Z value and the parameter address are not rewritten, but the polygon parameter address value that has been stored previously is transferred to the horizontal interpolation calculation device 33.

From the parameter calculation device 31, the X value (XL) on the left side, the Z value (ZL) on the left side, the difference between the X values on the left side and the right side (DXY), the fine difference between Z values (DDZ), and the polygon parameter memory address are sent. It is given to the parameter calculation device I / F 34b of the hidden surface processing device 34, and each data is transferred from the parameter calculation device I / F 34b to the scan line hidden surface processing device 34a.

The scan line hidden surface processing device 34a uses the Z
The hidden surface processing based on the buffer method is performed. In order to perform the hidden surface processing for each dot of the scan line, as shown in the block diagram of FIG. Then, the number (n + 1) of horizontal dot hidden surface processing devices 34-0 to 34-34 corresponding to the horizontal resolution.
-N is provided.

Then, the data received from the parameter arithmetic unit 31 is sequentially subjected to the pipeline processing from the 0 address of the continuous pipeline of the horizontal dot hidden surface processing units 34-0 to 34-n of the respective dot levels of the scan line. Performs high-speed hidden surface processing. The horizontal dot hidden surface processing devices 34-0 to 34-n correspond to each dot of the scan line and perform the hidden surface processing of the dot by the Z buffer method.

FIG. 11 shows an embodiment of the horizontal dot hidden surface processing apparatus. Depth distance data (Z) corresponding to each dot on the scan line transferred from the parameter calculation device 31.
The value) is stored in the depth register 341 via the parameter arithmetic unit I / F 34b. Parameter calculation device I /
The address of the polygon parameter memory 40 corresponding to each dot on the scan line transferred from F34b is stored in the parameter address register 342.

One of the inputs of the comparator 344 is the value of the state of the A bus, that is, the Z value of the polygon given through the parameter arithmetic unit I / F 34b, and the other input is
Depth register 341 via multiplexer 343
Either the previous Z value stored in or the X value corresponding to the dot, that is, the X address value is given. In this multiplexer 343, the comparator 344 uses the Z of another polygon.
When the value is compared with the value, the value of the depth register 341 is sent to the comparator 344, and when it is checked whether another polygon hits this dot, the X address value is sent to the comparator 344. Then, the comparator 344 determines whether or not the polygon flowing through the pipeline hits that dot, and if so, the Z value of that polygon is the Z value of the polygon stored in the parameter address register 342.
With respect to the value, it is determined which one is closer, that is, which Z value is smaller, and the controller 352 is notified of the result.

The adder 346 calculates the Z value and the difference D between the Z values.
By adding DZ, digital differential analysis (DD
A) is performed to obtain the Z value of the next dot, or the difference between the X values of the left side and the right side, and 1 is subtracted from DXY to obtain the right side value of X of the polygon. Therefore, the adder 34
DDZ and DXY are given to one of 6 from the B bus,
To the other side, the value of Z or "-1" of the state of the A bus is given from the multiplexer 345. Multiplexer 345
Sends the value Z of state 2 of the A bus to the adder 346 when the adder 346 calculates the Z value, and sends the value "-1" to the adder 346 when the DXY operation is performed.

In the multiplexer 348, the adder 346 is D
When the XY operation is performed, the output of the adder 346 is sent to the B bus pipeline register 350, and at other times, the value of the B bus is sent to the B bus pipeline register 350.

The adder 346 of the multiplexer 347 is Z
When the value is updated, the output of the adder 346 is sent to the A bus pipeline register 349, and at other times, the value of the A bus is sent to the A bus pipeline register 349. This A
The value of the bus pipeline register 349 is transferred to the horizontal dot hidden surface processing device 34 at the next stage.

The value of the bus of the B bus pipeline register 350 is transferred to the horizontal dot hidden surface processing device 34 at the next stage. The value of the C bus pipeline register 351 is transferred to the horizontal dot hidden surface processing device 34 at the next stage. The controller 352 is driven by an active signal received from the E bus, and if it is active, operates the horizontal dot hidden surface processing device according to the flow shown in FIG. 22, and if it is not active, does not drive the E bus pipeline flip-flop. The E bus signal indicating whether or not the controller 352 is active is stored in 353, and the E bus signal is transferred to the next horizontal dot hidden surface processing device 34.

FIG. 12 shows the data flow of each bus. The operation of the hidden surface processing apparatus will be described with reference to FIGS. 11 and 12.

First, as the right end point information of the scan line, the difference DXY between the right side and the left side is given from the B bus instead of the coordinate value. That is, the coordinates (XL, ZL) on the left side, DXY, and the fine difference value DDZ of the Z coordinate per unit dot are A bus and B bus from the left side of the horizontal dot hidden surface processing device 34 having the array structure shown in FIG. It is input separately. The control information (E bus signal stored in the flip-flop 353) includes information indicating that the dot is within the range. As shown in FIG. 12, the data of each bus is given in a time division manner.

The operation of each horizontal dot hidden surface processing device 34 will be described with reference to FIGS. 11 and 12. In the first timing state, XL is given to the A bus, DXY is given to the B bus, Z value is given to the A bus and DDZ is given to the B bus at the second timing, and both timings of state 1 and state 2 are given to the C bus. Gives the parameter address. When the comparator 344 is negative, this dot has entered the polygon existing range at this pixel position, and the multiplexer 345 gives the data “−1” to the adder 346. DXY is given to the other side of the adder 346 from the B bus. When the dot is within the polygon existing range, the multiplexer 347 outputs the output of the adder 346 to the register 349.

In the second state, if the dot is within the existing range, Z of A bus and DDZ of B bus are added and the output of adder 346 is output to B bus. For example, the data is output to the B bus without being updated. When the result of comparing the data ZA stored in the A data Z depth register 341 on the A bus with the comparator 344 in the state where the dots are within the existing range is Z <ZA,
The data in the depth register 341 is rewritten, and the parameter address register 342 stores the Z and C bus polygon parameter memory addresses, respectively.

The horizontal dot hidden surface processing device 34 shown in FIG.
At −0, 1 is subtracted and 0 is obtained, so that the Z value is not updated and compared with ZA. In the horizontal dot hidden surface processing device 34-1, 1 is further subtracted and becomes negative.
Y is output to the A bus, and the subsequent horizontal dot hidden surface processing device performs subtraction on DXY. The horizontal dot hidden surface processing device 34-1 further sets the flip-flop 353 to 1 and updates the Z value and compares it with ZA. In the hidden surface processing device in the latter stage, DXY is sequentially subtracted by 1, and the process is continued until the result becomes negative. When it becomes negative, flip-flop 35
3 is returned to 0, the remaining hidden surface processing device updates Z value, Z
No comparison with A is made.

FIG. 13 shows the scan line hidden surface processing device 34.
The time chart of the pipeline of a is shown. P1-1 is polygon state 1, P1-2 is polygon 1 state 2,
P2-1 is the state 1 of polygon 2. The dot address value on the horizontal line received from the horizontal interpolation calculation device 33 is sent to the scan line hidden surface processing device 34a, and the horizontal dot hidden surface processing device 34 corresponding to the next dot transfers the value of the parameter address register 342 to the D bus. And sends the polygon parameter memory address to the horizontal interpolation calculation device 33. Each of these devices is a controller 4
Controlled by 1.

Next, referring to FIG. 14, the horizontal interpolation calculation device 33
Will be explained. The hidden surface processing device 34 gives the address of the polygon parameter memory 40 for each dot to the parameter input device 33a. The parameter input device 33a reads the parameter stored at that address from the polygon parameter memory 40 according to the input polygon parameter memory address, and the interpolation mapping operation device 33
transfer to b.

The interpolating mapping arithmetic unit 33b receives the parameter input unit 33a from the polygon parameter memory 40.
X of the visual field coordinates for each polygon end point transferred via
Y and Z values and mapping pattern memory addresses MX and M
Y and a flag indicating the direction of mapping are received, and the mapping pattern memory address (MX, MY) of the dot of the X address of the scan line currently being processed is calculated,
Transfer to the frame memory controller 33c.

The frame memory controller 33c accesses the mapping pattern memory 7 by the mapping pattern memory address (MX, MY) obtained by the interpolation mapping arithmetic unit 33b, so that the dot of the X address of the scan line which is currently processed. R, G,
The B or LUT value is obtained and written in the frame memory 8. Each of these devices is controlled by the controller 42, and this controller 42 operates according to the flowchart shown in FIG.

FIG. 15 shows an interpolation mapping calculation device 33b.
It is a block diagram showing a specific example of. The interpolation mapping calculation device 33b includes a visual field coordinate X, Y, Z value generation device 331 and a mapping calculation device 332. The visual field coordinate X, Y, Z value generating device 331 receives the Z value converted into the normalized coordinate received from the hidden surface processing device 34,
This Z value is converted into the Z value of the visual field coordinates, and the screen X, Y
The X and Y values of the visual field coordinates are calculated from the values and the screen distance,
It is transferred to the mapping arithmetic unit 332.

The mapping calculation unit 332 receives the visual field coordinates X, Y, Z values from the visual field coordinate X, Y, Z value generation unit 331, and the polygon end points read from the polygon parameter memory via the parameter input unit 339. The X, Y and Z values of the visual field coordinates and the mapping direction, the reciprocal value 1 / ab of the outer product of the two sides on the visual field coordinates and the mapping pattern memory addresses MX and MY of the polygon end points are received,
Interpolation mapping calculation is performed and the mapping pattern memory addresses MX and MY of each dot are transferred to the frame memory controller 33c.

FIG. 16 shows a block diagram of a visual field coordinate X, Y, Z value generation device. Visual field coordinate X, Y, Z value generation device 33
1 is a visual field coordinate standard value generating device 331a and visual field coordinates X and Y
And a value generation device 331b. The visual field coordinate Z value generation device 331a outputs the normalized Z value by the conversion calculation of the following Equation 8. As the Z value normalization process, the one shown in the following formula 7 is known.

[0118]

(7) A = CZMAX / (VLEN * (CZMAX-
C = CZMIN)) B = CZMAX * CZMIN / (VLEN * (CZMA
X-CZMIN)) Z = (A * WZ-B) * VLEN / EZ where CZMAX is the maximum Z value, CZMIN is the minimum Z value, and VLEN is the distance from the screen.

By this processing, the visual field coordinate value ZE is converted to the normalized coordinate value IZ. Therefore, the visual field coordinate Z value generation device 331a performs the inverse transformation of the above equation 7 from the SZ value interpolated by the hidden surface processing device 34, and obtains the Z value of the visual field coordinate according to the following equation 8.

[0120]

[Equation 8] A2 = (CZMAX / (VLEN * (CZMA
X-CZMIN)) * VLEN B2 = CZMAX * CEMIN / (VLEN * (CZM
AX-CZMIN)) * (-VLEN) EZ = B2 / (IZ-A2)

The visual field coordinate Z value device 331a performs a calculation process so as to obtain the Z value of the visual field coordinates by the calculation of the above equation 8.

FIG. 17 shows a view coordinate Z value generation device 331a.
It is a block diagram showing a specific example of. The Z value of the normalized coordinates sent from the hidden surface processing device 34 is stored in the register 3
51 is stored in the register 51 and is stored in the register 352. The respective data is given to the subtractor 353 from both the registers 351 and 352, and 1Z
-A2 is calculated. The parameter of B2 is stored in the register 354, and this value is given to one of the dividers 355.

Field-of-view coordinate X, Y, Z value generator 331b
Is for calculating the projected X and Y values by the conversion operation of the following mathematical formula 10. As the projection process, the one shown in the following formula 9 is known.

[0124]

SX = EX * VLEN / EZ + CX where CX is the X value at the center of the screen.

Here, the above equation 9 is inversely transformed, and X and Y are calculated based on the following equation 10.

[0126]

EX = (SX−CX) * EZ * (1 / VLEN)

FIG. 18 shows a visual field coordinate X, Y value generating device 33.
It is a block diagram which shows the specific Example of 1b.

The CX and CY parameters of the above formula are stored in the register 361, and the CX and CY parameters from this register 361 are stored.
The CY parameter is given to the subtractor 363. The X value and the Y value of the screen coordinate are counted by the counter 362. The counter 362 counts up in synchronization with the entire system. The X value and the Y value of the screen coordinates are given from the counter 362 to the subtractor 363, and the subtractor 363 calculates SX-CX and SY-CY. The subtraction result from the subtractor 363 is given to the multiplier 364. EZ is given to the multiplier 364 from the visual field coordinate Z value generation device 331 a, and (SX−CX) × E
The calculation of Z, (SY-CY) * EZ is performed. On the other hand, the register 365 stores the 1 / VLEN parameter of the above equation, and 1 / VLEN is applied to the multiplier 366 from this register. This multiplier 366 has a multiplier 3
The multiplication result from 64 is given, and the visual field coordinates EX and EY are given from the multiplier 366 to the register 367.

FIG. 19 is a block diagram showing a concrete example of the mapping arithmetic unit 332.

The register 371 stores the visual field coordinates E of the polygon end points read from the polygon parameter memory 40.
X1, EX2, and EX3 are stored, and the register 372 stores the visual field coordinate X value (EX) received from the visual field coordinate X, Y, and Z value generation device 331. EX1, EX2, and EX3 stored in the register 371 are multiplexers 37
One of the inputs is applied to the subtractor 374 via the input terminal 3. The EX stored in the register 372 is applied to the other input of the subtractor 374. EX-EX is performed by the subtractor 373.
1, EX-EX2, EX-EX3 are calculated, and as a result, aX, bX, cX are stored in the register 375. From the register 375, aX, bX, and cX are given to the multiplexer 393, the multiplexer 402, and the multiplexer 411 via the multiplexer 376.

The register 377 stores the visual field coordinates E of the polygon end points read from the polygon parameter memory 40.
Y1, EY2, EY3 are stored, and the register 378 stores the visual field coordinate Y value (EY) received from the visual field coordinate X, Y, Z value generating device 331. EY1, EY2, and EY3 stored in the register 377 are multiplexers 37
It is given to one input to the subtractor 380 via 9. The EY stored in the register 378 is applied to the other input of the subtractor 380. This subtractor 380 produces an EY-EY
1, EY-EY2, EY-EY3 are calculated, and as a result, aY, bY, and cY are stored in the register 381. From this register 381, aY, bY, and cY are transmitted via multiplexer 382 to multiplexers 393, 4
20, 402, 421, 411, 422.

The register 383 stores the visual field coordinates E of the polygon end points read from the polygon parameter memory 40.
Z1, EZ2, EZ3 are stored, and the visual field coordinate Z value (EZ) received from the visual field coordinate X, Y, Z value generating device 331 is stored in the register 384. EZ1, EZ2, EZ3 stored in the register 383 are multiplexers 38
One of the inputs is given to the subtractor 386 via the signal V.sub.5. The EZ stored in the register 384 is applied to the other input of the subtractor 386. EZ-EZ in this subtractor 386
1, EZ-EZ2, EZ-EZ3 are calculated, and as a result, aZ, bZ, cZ are stored in the register 387. From this register 387, aZ, bZ, and cZ are transmitted via multiplexer 388 to multiplexers 393, 4
20, 402, 421, 422.

When the mapping direction is XZ, first,
One input of the multiplier 389 has a multiplexer 376,
CX is given from the register 375 via 393, bZ is given to the other input from the register 387 via the multiplexers 388 and 420, and W1K = bZ × cX
The calculation result W1K is stored in the register 390. Further, bX is given to one input of the multiplier 389 from the register 375 via the multiplexers 376 and 393, and the other input is input to the multiplexer 38.
CZ is given from the register 387 via 8, 420, W2K = bX × cZ is calculated, and the calculation result W
2K is stored in the register 391.

Then, aX is applied to one input of the multiplier 398 from the register 375 via the multiplexers 376 and 402, and the other input is input to the multiplexer 38.
CZ is given from the register 387 via 8 and 421, W1L = cZ × aX is calculated, and the calculation result W
1L is stored in the register 399. Furthermore, the multiplier 3
98 is provided with cX from the register 375 via the multiplexers 376 and 402 to one input of 98 and the register 3 via the multiplexers 388 and 421 to the other input.
87 is supplied with aZ, the calculation of W2L = cX × aZ is performed, and the calculation result W2K is stored in the register 400.

Further, bX is given to one input of the multiplier 407 from the register 375 via the multiplexers 376 and 411, and the other input is input to the multiplexer 38.
AZ is given from the register 387 via 8, 422, W1M = aZ × bX is calculated, and the calculation result W
1M is stored in the register 408. Furthermore, the multiplier 4
07 is supplied with aX from the register 375 via the multiplexers 376 and 411, and the other input is supplied with the register 3 via multiplexers 388 and 422 to the other input.
BZ is given from 87, W2M = aX × bZ is calculated, and the calculation result W2M is stored in the register 409.

When the mapping direction is YZ, first,
One input of the multiplier 389 has a multiplexer 376,
CX is given from the register 375 via 393, bY is given to the other input from the register 381 via the multiplexers 382 and 420, and W1K = bY × cX
The calculation result W1K is stored in the register 390. Further, bX is given to one input of the multiplier 389 from the register 375 via the multiplexers 376 and 393, and the other input is input to the multiplexer 38.
CY is given from the register 381 via 2, 420, W2K = bX × cY is calculated, and the calculation result W
2K is stored in the register 391.

Then, aX is given from the register 375 to one input of the multiplier 398 via the multiplexers 376 and 402, and the multiplexer 38 is provided to the other input.
CY is given from the register 387 via 2, 421, W1L = cY × aX is calculated, and the calculation result W
1L is stored in the register 399. Furthermore, the multiplier 3
98 is provided with cX from the register 375 via the multiplexers 376 and 402 to one input of 98 and the register 3 via the multiplexers 382 and 421 to the other input.
87 gives aY, W2L = cX × aY is calculated, and the calculation result W2L is stored in the register 400.

Further, bX is given to one input of the multiplier 407 from the register 375 via the multiplexers 376 and 411, and the other input is input to the multiplexer 38.
AY is given from the register 381 via 2, 422, W1M = aY × bX is calculated, and the calculation result W
1M is stored in the register 408. Furthermore, the multiplier 4
07 is supplied with aX from the register 375 via the multiplexers 376 and 411, and the other input is supplied with the register 3 via multiplexers 382 and 422 to the other input.
87 is provided with bY, W2M = aX × bY is calculated, and the calculation result W2M is stored in the register 409.

When the mapping direction is XY, first,
One input of the multiplier 389 has a multiplexer 376,
BX is given from the register 375 via 393, and cY is given to the other input from the register 381 via the multiplexers 382 and 420, and W1K = bX × cY.
The calculation result W1K is stored in the register 390. Further, one input of the multiplier 389 is supplied with cX from the register 375 via the multiplexers 376 and 393, and the other input is provided with the multiplexer 38.
BY is given from the register 381 via 2, 420, W2K = bY × cX is calculated, and the calculation result W
2K is stored in the register 391.

Then, one input of the multiplier 398 is supplied with cX from the register 375 via the multiplexers 376 and 402, and the other input is supplied with the multiplexer 38.
AY is given from the register 387 via 2, 421, W1L = cX × aY is calculated, and the calculation result W
1L is stored in the register 399. Furthermore, the multiplier 3
AX is given to one input of 98 from the register 375 through the multiplexers 376 and 402, and the other input is given to the register 3 through the multiplexers 382 and 421.
87 is supplied with cY, W2L = aX × cY is calculated, and the calculation result W2L is stored in the register 400.

Also, one input of the multiplier 407 is supplied with aX from the register 375 via the multiplexers 376 and 411, and the other input is provided with the multiplexer 38.
BY is given from the register 381 via 2, 422, W1M = aX × bY is calculated, and the calculation result W
1M is stored in the register 408. Furthermore, the multiplier 4
07 is supplied with bX from the register 375 through the multiplexers 376 and 411 to one input of 07 and the register 3 through the multiplexers 382 and 422 to the other input.
87 is supplied with aY, W2M = aY × bX is calculated, and the calculation result W2M is stored in the register 409.

The subtractor 392 includes registers 390 and 39.
W1K and W2K stored in 1 are given respectively, and K
= W1K-W2K is calculated, the weighting coefficient K is calculated,
K is stored in the register 394.

The subtractor 401 includes registers 399 and 40.
W1L and W2L stored in 0 are given respectively, and L
= W1L-W2L, the weighting coefficient L is calculated,
L is stored in the register 403.

The subtractor 410 includes registers 408 and 40.
W1M and W2M stored in 9 are given respectively, and M
= W1M-W2M, the weighting factor M is calculated,
Store M in the register 412.

The register 395 stores the mapping addresses MX1 and MY1 of the end points, and the register 39
Mapping address MX1 of K and the end point stored in 4
MY1 is multiplied by the multiplier 396, and WK = K × MX1
Alternatively, the calculation result of WK = K × MY1 is stored in the register 397.

Further, the register 404 stores end point mapping addresses MX2 and MY2, and the L stored in the register 403 and the end point mapping address M.
X2 and MY2 are multiplied by the multiplier 405, and WL = L ×
The calculation result of MX2 or WL = L × MY2 is stored in the register 4
It is stored in 06.

Further, the register 413 stores the end point mapping addresses MX3 and MY3. The M stored in the register 412 and the end point mapping addresses MX3 and MY3 are multiplied by the multiplier 414, and WM = M
The calculation result of × MX3 or WM = M × MY3 is stored in the register 415.

The WK, WL and WM stored in the registers 397, 406 and 415 are added by the adder 416, and the addition result is multiplied by the reciprocal 1 / ab of the outer product of the vector ab stored in the register 417. Is multiplied by
The calculation of (WK + WL + WP) × 1 / ab is performed, the mapping addresses MX and MY in the X and Y directions are calculated and stored in the register 419. Each of these circuits is controlled by the controller 423, and the controller 423
Operates based on the flowchart shown in FIG.

The operations of the polygon outer shape processing device 20 and the polygon inner processing device 30 will be described according to the above circuit example based on the operation flows of FIGS.

The operation of the polygon outer shape processing device 20 will be described. First, the controller 25 reads out the number of polygons (P) from the screen memory 5, and then reads out the number of polygon end points to be processed.
1 (steps S1 and S2).

Then, the starting points (XS, YS, ZS) of each side are read from the screen memory 5 (step S3), and the address of the screen memory 5 is incremented. Then, the end points (XE, YE, ZE) of each side are read from the screen memory 5 (step S).
4). A polygon extraction device 21 calculates a direction vector from the read start point (XS, YS) and end point (XE, YE) of the end point, determines the left side or the right side based on the side vector, and sets the polygon direction (DIR). Yes (step S5).

Then, the X parameter calculation device 22a of the polygon outline shape processing device 20 calculates the X parameter (step S6). Data of YE and YS from the screen memory 5 is given to the subtractor of the X parameter calculation device 22a, and the distance DY between them is calculated.

Subsequently, the start point (XS) and end point (XE) data is given from the screen memory 5 to the subtractor of the X parameter arithmetic unit 22a, and the subtraction result (XE-XS) from this subtractor is supplied to the divider. To be done.

In this divider, (XE-XS) / DY is divided to calculate the X parameter.

Subsequently, an interpolation calculation is performed by the X vertical interpolation device 23a using this X parameter (step S).
9). This value (DDX) is given to the adder of the interpolation calculation circuit. Interpolation of X + DDX is performed by this adder,
This value is written in the register (step S9).

DY is given to this divider from the subtractor 62 of the difference circuit, and DY is divided by the above calculation result to calculate a fine difference value.

In this fine difference calculation circuit, DZ = (ZE-
ZS) / DY is calculated and parameters are calculated. The calculation result is supplied to the adders of the vertical interpolation device 23a and the Z value interpolation device 23b.

In steps S10 and S11, the interpolation device 2
The outputs from the fine difference calculation circuit are added to the adders 3a and 23b,
Addition is performed with the previous data set in the register, and Z = Z + DZ is calculated.

This value is given to the register, and the value of this register is written in the polygon edge memory 6 as the address data of the value of the register being the Z value.

The polygon edge memory 6 stores a left side address, a right side address, a polygon memory address, a Z side left side address and a right side address of a polygon for each Y address.

In step S11, Y + DIR is calculated, that is, "+1" is calculated when the polygon faces downward, and "-1" is calculated when the polygon faces upward, and the process proceeds to step S14. Then, in step S12, Y ≠ YE is determined, and Y ≠ Y
In the case of E, the process returns to step S9, the above-described operation is repeated, and when Y = YE, the process proceeds to step S13.

In step S13, it is determined whether or not the sides of all the polygons are finished. If not,
Returning to step S3, the above-described operation is repeated.

When all the sides of the polygon are completed, the process proceeds to step S14, and it is determined in step S14 whether all the processes of the polygon are completed. If all the processes of the polygon are not completed, Returning to step S2, the above-mentioned operation is repeated. Then, when it is determined that all the polygons have been processed, the outer shape processing operation ends.

Next, the internal polygon processing device 30 will be described. First, reading from the polygon edge memory 6 will be described with reference to FIG.

First, the internal polygon processing apparatus 30 initializes the Y address (step S21), reads the number of polygons for each Y address of the scan line, and (step S2).
2), go to step S23.

In step S23, the outline of the polygon between the left sides of the two sides facing each other for each Y address is shown. 2 X
The left side (XL) and the Z value address (ZL) are read from the polygon edge memory 6.

Then, the parameter calculation device 31 performs parameter calculation (step S25). The subtractor is supplied with the XR and XL data from the polygon edge memory 6, and the distance DX between them is calculated. This DX is supplied to the fine difference calculation circuit.

The polygon edge memory 6 and the left side (ZL) and right side (ZR) data of the Z address are given to the subtractor in the fine difference calculation circuit, and the subtraction result ZL-ZR from this subtractor is given. Supplied to the divider.

This divider performs division of (ZR-ZL) / DX (step S26), and this value (DDZ) is given to the hidden surface processing device 34 (step S27). In step S28, it is determined whether or not all polygons on the scan line have been processed. If the processing has not been completed, the process returns to step S23 to repeat the above operation.

Further, when one scan line, that is, the number of polygons at the Y address is completed, the process proceeds to step S29, the Y address is incremented at step S29, and the processes corresponding to all the Y addresses are completed.
That is, the above-described operation is repeated until the Y address becomes larger than the vertical resolution (step S30), and the parameter calculation operation ends when the processing corresponding to all the addresses ends.

Next, the operation of the horizontal dot hidden surface processing device 34-n will be described with reference to the operation flow of FIG. First, it is determined whether or not the dot exists within the polygon. That is, the scan line address of the Y address (for each scan line) is given from the polygon edge memory memory controller 31a, this XL is compared with the X address L corresponding to the dot, and whether the value of DXY is 0 or not is determined. It is determined (step S71). If XL is larger than the X address and DXY is not 0, the dot exists within the polygon range, so the process proceeds to step S72. If not, the hidden surface processing operation is repeated.

In step S72, the depth register 34
The Z value (ZA) stored in 1 is compared with the Z value that has just been read, and if the Z value of the depth register 341 is larger, that is, if the dot that has just been read is on the front side, step S73. If the Z value is small, the process proceeds to step S74. In step S73, the polygon parameter address is stored in the polygon parameter address register 342, and the flow advances to step S74. In step S74, the operation of DXY = DXY-1 is performed, and step S75
After proceeding to step S75, the Z value is interpolated in step S75, that is, Z = DDZ + Z is calculated.
The hidden surface processing operation ends.

Next, the data read operation from the polygon parameter memory 40 in the polygon internal processing device 30 will be mainly described with reference to the operation flow chart of FIG. First, after initializing the Y address and the X address (steps S41 and S42), the hidden surface processing device 34 performs X
The polygon parameter address of the address is read (step S43).

Subsequently, in step S44, it is determined whether or not a polygon exists at the X address. If there is no polygon, the process proceeds to step S50. If there is a polygon, the process proceeds to step S45. . In step 45, parameters are calculated by the parameter calculation device, and the process proceeds to step S46. In step S46, the polygon parameter memory 40 is set in accordance with the polygon parameter address corresponding to the X address given by the hidden surface processing device 34.
Then, the polygon parameters are read out, and the process proceeds to step S47.

In step S47, the mapping pattern memory 7 is read according to the parameters read from the polygon parameter memory 40 by the horizontal interpolation computing device 33.
Address (MX, MY) is calculated, and the process proceeds to step S48.

In step S48, the mapping pattern memory 7 is accessed according to the calculated address, and R, G, B or L is read from the mapping pattern memory 7.
Color information such as the UT value is read, and the process proceeds to step S49.

In step S49, the color information is written in the frame memory 8, and the process proceeds to step S50. In step S50, the X address is incremented by 1, and the process proceeds to step S51.

In step S51, the X address is compared with the horizontal resolution. If the X address is smaller than the horizontal resolution, the process returns to step S43 and the above-described operation is repeated. When the X address becomes larger than the horizontal resolution, the process proceeds to step S52, and in step S52, the hidden surface processing device 34
Is initialized and the process proceeds to step S53.

At step S53, the Y address is incremented by one, and the process proceeds to step S54 and step S5.
At 4, the Y address and vertical resolution are compared. If the Y address is smaller than the vertical resolution, the process returns to step S42, and the above-described operation is repeated. When the Y address becomes larger than the vertical resolution, the polygon internal processing operation ends.

Next, the operation of the interpolation mapping arithmetic unit 33b will be described with reference to the flowchart of FIG.

From the geometrical transformation device 2, the polygon parameter memory 40 is provided with flags corresponding to the mapping directions XZ, YZ, and XY in accordance with the normal vector direction. In this embodiment, when the flag is "0", the mapping direction is the XZ direction, when the flag is "1", the mapping direction is the YZ direction, and when the flag is "2",
The mapping direction is the XY direction.

First, in step S80, the flag (FL) read from the polygon parameter memory 40 is determined. When FL = 0, the process proceeds to step S82, and the interpolation mapping operation is performed when the mapping direction is XZ. Done. If FL = 0 is not satisfied, the process proceeds to step S81,
It is determined whether or not FL = 1. When FL = 1, the process proceeds to step S85, and interpolation mapping calculation is performed when the mapping direction is YZ. If it is determined in step S81 that FL = 1 is not satisfied, the process proceeds to step S87, and interpolation mapping calculation is performed when the mapping direction is XY.

In step S82, aX = EX-EX1, aZ = EZ-EZ1 bX = EX-EX2, bZ = EZ-EZ2 cX = EX-EX3, cZ = EZ-EZ3 are calculated, and the process proceeds to step S83.

In step S83, W1K = bZ × cX, W2K = bX × cZ W1L = cZ × aX, W2L = cX × aZ W1M = aZ × bX, W2M = aX × bZ are calculated, and the process proceeds to step S84.

When the mapping direction is YZ, the YZ interpolation mapping operation is performed in step S85.

In step S85, aY = EY-EY1, aZ = EZ-EZ1 bY = EY-EY2, bZ = EZ-EZ2 cY = EY-EY3, cZ = EZ-EZ3 are calculated, and the process proceeds to step S86.

In step S86, W1K = bY × cX, W2K = bX × cY W1L = cY × aX, W2L = cX × aY W1M = aY × bX, W2M = aX × bY are calculated, and the flow advances to step S84.

When the mapping direction is XY, in step S87, XY interpolation mapping calculation is performed.

At step S87, aX = EX-EX1, aY = EY-EY1 aY = EY-EY2, bX = EX-EX2 cY = EX-EX3, cY = EY-EY3 are calculated, and the routine proceeds to step S88.

In step S88, W1K = bX × cY, W2K = bY × cX W1L = cX × aY, W2L = cY × aX W1M = aX × bY, W2M = aY × bX are calculated, and the flow advances to step S84.

In step S84, the weighting factors K, L, M
To calculate K = W1K-W2K, L = W1L-W2L, M = W1M
-W2M is calculated, and the process proceeds to step S89.

In step S89, in order to calculate the weighting coefficient for the mapping in the X direction, WK = K × MX1, WL = L × MX2, WM = M × MX
3 is calculated and the process proceeds to step S90.

In step S90, the WX = (WX = (W
K + WL + WM) × 1 / ab is calculated, and step S
Proceed to 91.

In step S91, the weighting coefficient for the mapping in the Y direction is calculated. Therefore, WK = K × MY1, WL = L × MY2, WM = M × MY
3 is calculated and the process proceeds to step S92.

In step S92, WY = (WK in order to obtain the mapping address in the Y direction in consideration of the weight coefficient.
+ WL + WM) × 1 / ab is calculated, and this operation ends.

[0196]

As described above, according to the present invention, the projection direction of mapping is XZ or YZ from the direction of the normal vector of the three-dimensional polygon on the visual field coordinates.
Alternatively, it is determined whether it is XY. One of them is selected, and interpolation mapping processing for obtaining the two-dimensional mapping pattern coordinates from only the two-dimensional element is performed. As a result, by performing a square root calculation that requires a large amount of hardware, the calculation process can be largely omitted, and simple perspective texture mapping can be processed in real time at low cost.

Further, according to the present invention, the depth information indicating the depth state of each polygon is provided, the depth information is changed in accordance with the outer shape of the polygon, the interpolation is performed, and the Z buffer method is used. By the hidden surface processing described above, the mapping pattern address of the polygon located at the forefront at the compared pixel position is calculated by interpolation mapping calculation. By accessing the mapping pattern memory in accordance with the calculated mapping pattern memory address, the R, G, B or luminance (LUT) values of each pixel are sequentially read out based on the data stored in the mapping pattern memory, thereby achieving high speed. The mapped image can be displayed on the display device.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a stereoscopic image display device of the present invention.

FIG. 2 is a block diagram showing a configuration of a polygon outline shape processing device used in the present invention.

FIG. 3 is a block diagram showing a configuration of an X parameter calculation device in the polygon outer shape processing device.

FIG. 4 is a block diagram showing a configuration of a Z parameter calculation device in the polygon outer shape processing device.

FIG. 5 is a block diagram showing a configuration of an X vertical interpolation device in the polygon outline shape processing device.

FIG. 6 is a block diagram showing a configuration of a Z vertical interpolation device in the polygon outline shape processing device.

FIG. 7 is a block diagram showing a configuration of a parameter calculation device of the polygon internal processing device used in the present invention.

FIG. 8 is a block diagram showing a specific configuration of a hidden surface processing parameter calculation device of the parameter calculation device.

FIG. 9 is a block diagram showing a configuration of a hidden surface processing device of the parameter calculation device.

FIG. 10 is a block diagram showing a configuration of a scan line hidden surface processing apparatus which constitutes the hidden surface processing apparatus.

FIG. 11 is a block diagram showing a specific configuration example of a horizontal dot hidden surface processing apparatus which constitutes the scan line hidden surface processing apparatus.

FIG. 12 is a timing chart showing a data transfer state to the horizontal dot hidden surface processing apparatus.

FIG. 13 is a schematic diagram showing processing timing of the scan line hidden surface processing apparatus.

FIG. 14 is a block diagram showing a configuration example of a horizontal interpolation calculation device of an internal drawing processing device used in the present invention.

FIG. 15 is a block diagram showing a specific example of an interpolation mapping calculation device of the horizontal interpolation calculation device.

FIG. 16 is a visual field coordinate X of the interpolation mapping calculation device,
It is a block diagram which shows the specific Example of a Y, Z value generator.

FIG. 17 is a block diagram showing a specific example of a visual field coordinate Z value generation device of the interpolation mapping calculation device.

FIG. 18 is a visual field coordinate X of the interpolation mapping calculation device,
It is a block diagram which shows the specific Example of a Y value production | generation apparatus.

FIG. 19 is a block diagram showing a specific example of the interpolation mapping calculation device.

FIG. 20 is a flowchart showing the operation of the polygon outline shape processing device of the present invention.

FIG. 21 is a flowchart showing the operation of the polygon internal processing device of the present invention.

FIG. 22 is a flowchart showing the operation of the polygon internal processing device of the present invention.

FIG. 23 is a flowchart showing the hidden surface processing operation of the polygon internal processing device of the present invention.

FIG. 24 is a flowchart showing an interpolation mapping processing operation of the polygon internal processing device of the present invention.

FIG. 25 is a schematic diagram showing a screen memory used in the present invention.

FIG. 26 is a schematic diagram showing a polygon parameter memory used in the present invention.

FIG. 27 is a schematic diagram showing a polygon edge memory used in the present invention.

[Fig. 28] Fig. 28 is a diagram illustrating a relationship in a side vector direction of a polygon.

FIG. 29 is a diagram showing a relationship between a direction vector of a polygon and a side.

FIG. 30 is a schematic diagram showing an example of a mapping pattern memory.

FIG. 31 is an explanatory diagram illustrating a conventional interpolation mapping processing method.

FIG. 32 is an explanatory diagram illustrating a conventional interpolation mapping processing method.

FIG. 33 is a flowchart showing the operation of a conventional interpolation mapping processing method.

[Explanation of symbols]

 1 world memory 2 geometric conversion device 5 screen memory 7 mapping pattern memory 8 frame memory 9 CRT 20 polygon outer shape processing device 30 polygon internal processing device 40 polygon parameter memory

Claims (2)

[Claims] 1. A storage means for storing the end point information of a triangular pattern, the end point information of a mapping pattern memory for storing a texture image, and the depth information of the polygon, and coordinate conversion of each end point information from the storage means. , Conversion means for converting the screen end point information and selecting the projection direction of the mapping polygon according to the direction of the normal vector on the polygon visual field coordinates, and the cross product value of the vectors of the two sides of the polygon based on the projection direction of the mapping polygon. A means for calculating the reciprocal, a polygonal parameter storage means for storing the reciprocal of the visual field coordinate value of the end point of the polygon, the end point information of the mapping pattern memory, the projection direction of the mapping polygon and the cross product of the vectors of the two sides, and the conversion Screen end point information from the means, polygon depth information and the above polygon parameters. Screen storage means for storing the address value of the meter storage means, and address information of the polygon outer shape and polygon depth information based on the screen end point information from the conversion means, and the address value of the polygon parameter storage means for each scan line A contour line processing unit for converting into contour information, a unit for computing the information between the two opposite sides calculated by the contour processing unit, and calculating the displacement of the X address and the displacement of the depth information, and the scan line. Means for determining whether or not the pixel position corresponding to each pixel in the polygon exists within the range of the polygon, and the depth information of the foremost pixel and the depth information of the polygon at the pixel position are compared and compared. And the depth of the polygon that always rewrites the depth information of the Means for calculating the depth information of the adjacent pixel position by adding the displacement of the depth information to the pixel information, and at least one X address between two opposing sides of the polygon existing at the front at the compared pixel position and the polygon. A means for calculating an address value of the parameter storage means and a weighting coefficient of three vertices of the polygon based on the information of the polygon parameter storage means read out by accessing the calculated address value of the polygon parameter storage means, A stereoscopic image processing apparatus comprising: means for calculating an address value of a mapping pattern memory; and means for accessing the mapping pattern memory based on the calculated mapping pattern address and transferring image data to a display means. 2. The stereoscopic image processing apparatus according to claim 1, wherein the projection direction of the mapping polygon is selected according to the magnitude of the scalar quantity of the normal vector on the polygon visual field coordinates.