JP2011138445A - Transparent object display circuit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide: a drawing means of videos involving refraction reflected on a transparent object such as glass; and a logical circuit implementation of the means, concerning a computer graphic drawing technology. <P>SOLUTION: An environmental cube surrounding the transparent object and an object space are provided, and the object space is further divided into smaller spaces. A refractive vector T is determined by a refraction calculation circuit 40 from a visual line vector V and a surface-normal N at the drawing point of the transparent object, and a linear expression is determined at a linear expression generation circuit 41. A crossing test of the line and the divided spaces is performed at the circuit 41 and a circuit 42, and if they form a surface crossing space, predetermined polygon data is read out. The polygon data is used to determine a crossing point coordinate value in the circuit 41 and a crossing calculation circuit 45. The crossing point coordinate value and the surface-normal are given to the refractive calculation circuit 40, and a vector projecting from the object is determined and testing of the next crossing point and crossing calculation are repeated. A crossing point with the environmental cube is determined using a vector reaching the outlying of the object space in the circuit 41, and the brightness thereof is used as the brightness of the initial drawing point. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a computer graphics and a technical field of image drawing means with refraction projected on a transparent object such as glass and a logic circuit mounting of the means.

In computer graphics, the video expression projected on a transparent object is the brightness (video) of the point where the line-of-sight vector at the drawing point is incident on the object and refracted, and the line-of-sight vector that exits the object intersects the image ahead of it. And this can be expressed as the luminance on the object (drawing point). Conventionally, the ray tracing method has been used for high-definition video representation as a means to determine the luminance at the drawing point relatively accurately, but this method has a heavy computer load and is suitable for real-time processing. Absent. In the polygon rendering method that emphasizes real-time characteristics, a refraction vector that passes through the inside of an object is obtained from the normal of the object at the drawing point, the line-of-sight vector, and the refractive index of the object, and this vector is outside the object on the extension line. A method has been proposed in which the luminance of a point that intersects with the image is obtained, and only the first-order refraction on the first object is considered as the luminance of the drawing point. The intersection at which the refraction vector finally intersects is obtained by calculating an intersection with the cube plane by providing six environmental cubes in advance, storing images projected on the six planes with the cube center as the projection center. Since this method does not consider the second-order refraction when the refraction vector escapes from the transparent object, the image is significantly different from the actual situation. On the other hand, from the position of the intersection on the cube surface obtained by the first-order refraction of the refraction vector and the position of the intersection of the cube surface obtained by extending the drawing point coordinate value, a position approximately in the middle is set as a new intersection of the refraction vector. A method for reducing the error of first-order refraction has also been proposed. However, in any method, a more realistic image expression cannot be obtained unless the correct position where the refraction vector inside the object comes out of the object again and the secondary refraction action at that time are taken into consideration.
In the present invention, the position of the second-order refraction point where the refraction vector exits from the object is accurately obtained, the refraction at that point is calculated, and further suitable for the real time in which multi-order refraction occurring in an object having a complicated shape is possible. We propose an implementation circuit for drawing means and its algorithm.

  An object of the present invention is to implement an algorithm for executing drawing in consideration of multi-order refraction of a transparent object in real time and its hardware implementation.

  In drawing a transparent object defined by polygon data, a line-of-sight vector and a surface normal are interpolated from polygon vertex data as drawing point information. Further, since the transmittance of the transparent object is also defined in advance, a line-of-sight vector (hereinafter referred to as a refraction vector) inside the object can be obtained. The position where the refraction vector intersects and exits from the object is the most computationally intensive and has not been obtained by real-time processing with conventional techniques. In the present invention, drawing of a transparent object is divided into two processes: preprocessing and drawing processing. Since the pre-processing does not occur unless the object is deformed, the drawing speed is not affected. Based on this assumption, in the present invention, a rectangular parallelepiped space (hereinafter referred to as an object space) surrounding an object is defined, and this object space is divided into spaces of an arbitrary size. Each divided object space (hereinafter referred to as a divided space) includes a divided space where the object planes intersect each other, and a divided space having three states that are completely inside the object and completely outside the object. Each buffer having an address corresponding to the coordinate value of the divided space is provided to store the divided space information. That is, an address structure in which m, n, and p are simply arranged in a bit field may be used as long as they are mxnxp divisions (arbitrary values) with respect to the x, y, and z coordinates. Therefore, the buffer is composed of the number of addresses corresponding to the number of divisions and the number of words of a predetermined size per address. In the present invention, the divided space in the three states is stored in the buffer at the corresponding address as flag information of, for example, A outside, B inside, and C including the object plane.

  On the other hand, polygon information included in the space is stored in the divided space including the object plane. This stores the address for linking to the address of another buffer that stores polygon information in the corresponding address buffer, ignores if A and B, and reads the link address if there is a C flag. This is a method of reading the polygon information ahead. A plurality of polygons may be stored in one divided space, or the same polygon may be stored in a plurality of divided spaces when the same polygon straddles a plurality of divided spaces. The number of polygons per division space depends on the division size, and optimization can be achieved because space division is performed by preprocessing.

  During the execution of drawing, the refraction vector of the drawing point (primary refraction point) is obtained. In the present invention, a straight line expression is established from the drawing point coordinate value and the refraction vector, and only the divided space intersecting with the straight line is searched. Straight line stepping is assigned to the linear expression of each of the three axes using the coordinate axis in which the straight line is the long axis (the axis with the greatest inclination of the straight line) in the three-dimensional coordinate system and the side length of the divided space as a unit. It is sufficient to test only the partition space that contains the points obtained. This test is to read and check the A, B, and C flags of [004] from the buffer, and if A and B, skip and further advance to detect the flag C. If the flag C is detected, polygon information is read from a predetermined link address, and intersection calculation between the straight line type and the polygon surface is performed. The intersection calculation between the straight line and the surface can be performed by a known mathematical process.

  In the present invention, attribute information at the vertices is defined together with the vertex data of the triangle as polygon information. They are the surface normal, the refractive index, the transmittance, the attenuation factor, and the texture address. When the triangular surface is a part of a polyhedron, the surface normal at each vertex is different. Therefore, the surface normal at the intersection is defined as 3 in each polygon vertex from the intersection coordinate value of the refraction vector and the polygon surface. A three-point interpolation method with two normals is used. This can be dealt with by the same interpolation method even when the transmittance and attenuation rate differ depending on the surface processing of the transparent object, and the aforementioned transmittance and attenuation rate are defined differently at the vertex of the polygon. On the other hand, when these are defined by the map method instead of the vertex attribute, there is a method of obtaining them from the texture coordinates of the intersection. Further, since the three-point interpolation method is obtained by four arithmetic operations in the same manner as the intersection calculation between a straight line and a surface, high-speed processing can be performed by circuitizing the interpolation processing.

  From the coordinate value at the intersection (secondary refraction point) of the refraction vector, the surface normal, and the refractive index, a refraction vector (hereinafter referred to as an output line-of-sight vector) that exits from the transparent object is obtained. Using this output point and the output line-of-sight vector, an intersection point with the environment cube surface further surrounding the transparent object is obtained, and the luminance of this surface is set as the luminance of the drawing point (first-order refraction point). The method of calculating the intersection with the environment cube surface from a predetermined point is a method known in the (reflection vector) position-dependent environment cube drawing technique, and is not included in the new scope of the present invention, so that details are omitted.

  In refraction calculation, total reflection may occur due to the relationship between the refractive index and the input line-of-sight vector or the refraction vector and the normal. If it is total reflection on the external surface of the object, the reflection vector is obtained from the line-of-sight vector and the surface normal, the total reflection point is set as a new start point, and the divided space is continuously tested by the same means as [006] to obtain the next intersection. . When the divided space becomes the outer divided space of the object space, a point intersecting with the environment cube surface on the extension line of the reflection vector is obtained. Similarly, in the case of total internal reflection, the divided space is continuously tested by means of [006] to obtain the next intersection. There may be a case in which total reflection is repeated inside and the environment cube is not reached. In this case, it is necessary to set an upper limit of the refraction order, and when it exceeds the upper limit, it is aborted and set to a predetermined brightness.

  When testing a straight line and a divided space in [006], two coordinate values obtained by substituting into a linear equation in two dimensions, and four divided spaces in three dimensions become a test space for a straight line and a polygon surface. If a C divided space is detected, an intersection (secondary refraction point) is obtained by the same means as [007] [008]. If there are no intersections, the linear equation is advanced until further intersections are detected. When an intersection is first detected for an object having a concave surface, the intersection with the environment cube proposed in [008] cannot be obtained immediately. In this case, it is necessary to obtain a linear equation for each refraction point and continue until the straight line reaches the outer surface of the object space.

  In the present invention, in the concave shape, the information at the drawing point (first-order refraction point) is saved, and at the same time, the output line-of-sight vector at the second-order refraction point is obtained, and the process of [006] is started again from the second-order refraction point. . This is executed until the outer surface of the object space of the transparent object is reached. As a result, for example, when the last nth-order refraction point and its nth-order output line-of-sight vector are obtained, the intersection calculation with the environmental cube is started using them. The stored information is used for final luminance determination.

  In refraction calculation, when the refraction vector that passes through the object or the output line-of-sight vector crosses the next surface and passes through the object again, if some opaque material is mixed or the object surface is painted In some cases. To represent them, each vector starting from the input viewpoint vector at every intersection is either a refraction or a reflection vector at each intersection (some reflections have neither specularity nor specularity) And test the next intersection in the determined vector direction.

In the present invention, the attribute [007] is defined as polygon information. Here, the attenuation rate is a luminance attenuation scale value that attenuates the luminance obtained from the environment cube by the passage length of the refraction vector in the object (distance from the first-order refraction to the intersection). The transmittance is high when the transmittance is high, and the transparency is strong. When the transmittance is low, the transmittance is opaque, and the texture luminance at the intersection is obtained from the texture coordinate value.
In this case, the refraction vector intersects the opaque surface, and further refraction / reflection processing is aborted. Therefore, the texture color becomes the color of the primary refraction point (drawing point), and the processing per drawing point ends even if the refraction vector is in the object space.

  According to the present invention, a multi-order refraction image that is transmitted through a transparent object can be drawn at a high speed, and the realism can be enhanced in a real-time computer graphics image.

  The means of the present invention is implemented in a graphics LSI as embedded software or a circuit, or is implemented in the form of IP (Intelligent Property).

  Examples of the present invention will be described below. FIG. 1 shows a space 11 surrounding a transparent object 12 by dividing a space surrounding a transparent object according to the present invention into a lattice shape. In determining the luminance at the drawing point P on the transparent object, the refraction (first-order refraction) at the object point P is obtained from the line-of-sight vector V of the viewpoint 10 and the surface normal Np of the drawing point, and the path 13 of the refraction vector T is obtained. obtain. The object space 11 is preliminarily divided into an arbitrary size and listed, and the divided space where the object plane 12 intersects is flagged as C, the space completely outside the object is A, and the complete interior of the object is B. . When the number of the divided space is expressed by an address (x, y) on the x and y axes, in the example of FIG. 1, the P point is at (1, 4) (2, 4), (3,4), Crosses the object plane at point Q of (5, 3) through (4, 3). The actual object is three-dimensional in three dimensions, and the address of the divided space is three axes including the z axis.

  The path 13 of the refraction vector T passes through the split space flag C at (1, 4) (4, 3) and (5, 3). Since this divided space is a flag indicating that there is an object surface, the intersection calculation with each polygon of the object surface registered in the list is performed. In this intersection calculation, a straight line is derived from the point P and the refraction vector, and a generally known intersection test between this straight line and the registered polygon plane is performed. When the intersection point Q is obtained by the intersection calculation, the refraction calculation (second-order refraction) at the point Q is performed to obtain the output line-of-sight vector 14 that goes out of the object. This vector 14 passes through (5, 3) and (6, 3), but the address (6, 3) is the outer space of the object space C. If there is no intersection plane at this address, the output line-of-sight vector 14 is the object. The calculation of the refraction intersection with the object is completed because it has gone outside. Thereafter, the luminance of the environment cube surface where the output line-of-sight vector 14 intersects is obtained by environment mapping including a transparent object, and the surface attribute effect is added to this to obtain the luminance of the point P, whereby a refraction image can be expressed.

  FIG. 2 shows a relationship for determining a divided space through which a refraction vector passes according to the present invention. In the object space, there are many divided spaces that are unrelated to the passband of the refraction vector. In the intersection test, it is necessary to detect only the divided space concerned at high speed and test the stored flag. In the present invention, a straight line 20 is derived from the refraction vector T and the object entrance point P0 at the viewpoint. The linear expression is an expression for each axis having the major axis (the axis with the largest inclination of the straight line) among the x, y, and z axes, and the unit width (x1−x0) of the divided space is Substitute and advance a straight line. Referring to the two-dimensional coordinates in FIG. 2, the straight line 20 is inclined with respect to the x-axis, and the linear formula is given by yi = (Δy / Δx) (xi−x0) + y0 with the x-axis as a base axis. Here, yi is the value of y at point xi. Further, (x0, y0) is the coordinate value of the starting point P0. The step value of Xi is the unit length of the divided space. In FIG. 2, the y-axis intersects at C0 at x1, intersects C1 at x2, and intersects C2 at x3. The divided spaces shared by C0 and x1 are the left and right sides. Similarly, the divided spaces (shaded portions in FIG. 2) where C1 and C2 also share the respective intersections are regions through which the straight line 20 passes. As long as the straight line does not move one step along the y-axis, the flag test for the previous area and the next area will be double, so even if there are two shared areas for each assignment, the test will Once. As described above, the area sharing the coordinate point obtained by substituting the long axis coordinate value can be easily detected. The target area flag is tested, and the straight line is updated until an object plane presence / absence flag is detected. This update ends when the outer space of the object space is reached.

  FIG. 3 shows a relationship diagram of multi-order refraction according to the present invention. The input line-of-sight vector V from the start point 30 enters the point P0 (current drawing point) on the surface of the object 31, and the refraction vector T0 inside the object is obtained using the surface normal N0 of the point P0, the transmittance, and the Fresnel equation. The intersection point Q0 is obtained through the intersection test process in the divided space, and the output line-of-sight vector T1 that has exited the object is obtained from the surface normal N1 and the transmittance at that point. Further, when the next intersection is obtained in the same manner as described above and the point P1 is detected, the refraction vector T2 inside the object is obtained again from the surface normal N2 and the transmittance at that point. Similarly, an output line-of-sight vector T3 is obtained through Q1 and the surface normal N3. Further, a divided space test is performed, but the object outline 32 is reached, so the intersection test and calculation are finished at this point. From the vector T3 and the last point Q1, the brightness of the point that intersects the cube of T3 (not shown in the figure) with respect to the environment cube further surrounding the object 31 is obtained, and this is set as the brightness of the point P0.

  The brightness acquired from the environmental cube can be attenuated in consideration of the attribute value of each of the intersections P0 and P1, or the translucency and scattering characteristics inside the object. This is obtained by storing attribute information for each intersection and performing a predetermined calculation (for example, alpha blending) using the stored attribute value when the environment cube luminance is obtained.

FIG. 4 shows a logic circuit configuration diagram of the present invention. The line-of-sight vector V and the surface normal N at the drawing point are added to the refraction calculation circuit 40. Here, the refractive belt T is obtained using a refractive index given in advance. The circuit is composed of Fresnel refraction. Further, the linear expression generation circuit 41 obtains a linear expression based on the drawing point coordinate value P (x, y, z), the T and the size of the known object space. In the linear formula, the step starts by substituting the division space unit value on the long axis side sequentially. The information output from the circuit 41 is a coordinate value at which the straight line intersects the lattice axis of the divided space, and becomes the number (address) of the divided space that shares this coordinate value. This address is added to the volume boundary and divided space intersection test circuit 42, and the divided space state buffer 43 is accessed to read the flag. The read flag tests in circuit 42 whether the partition space is a complete interior, exterior or object plane intersection. The circuit 41 is stepped up completely inside or outside. On the other hand, if it is a plane crossing space, the plane crossing polygon buffer 44 storing each polygon information described in [007] corresponding to the divided space is accessed. Then, predetermined polygon information is read out. The polygon information is usually a triangle, and is composed of coordinate values of three vertices constituting the polygon, surface normals, and the like. From the read polygon information, a straight line equation of the circuit 41 and an intersection calculation circuit 45 obtain an intersection coordinate value. In the case of a polygon that does not intersect, the polygon 45 is read out from the circuit 45 to 42. When the intersection point is obtained, the coordinate value and surface normal at the intersection point are given to the refraction calculation circuit 40 via the path A, and the vector that has escaped from the object is obtained, and the first refraction vector is already in the circuit 40. The same intersection test and intersection calculation are repeated again using the above. When the step of the straight line reaches the outer space of the object space in the circuit 41, the intersection test ends.
In order to speed up a series of these processes, it is possible to form a pipeline structure corresponding to the refraction order by combining the circuit 40 to the circuit 45 as one set.

  The circuit of the present invention is provided as an IP (Intelligent property) or mounted on a graphics processor LSI so that it can be used for a realistic CG video production or an amusement system.

“The relationship between the object space, the refraction vector, and the divided space according to the present invention is shown.” “The relationship between the line of sight or the refraction vector passing space according to the present invention is shown.” “Shows the multi-order refractive state according to the present invention.” “A transparent object display circuit according to the present invention is shown.”

FIG.
10 viewpoint 11 object space 12 transparent object 13 refraction vector 14 object external refraction vector
20 Refraction vector line 21 Divided space lattice Fig. 3
30 viewpoint 31 transparent object 32 object space
40 Refraction Calculation Circuit 41 Linear Generation Circuit 42 Volume Boundary and Division Space Intersection Test Circuit 43 Division Space State Buffer 44 Plane Intersection Polygon Buffer 45 Intersection Calculation Circuit

Claims (4)

  Regarding the drawing circuit of transparent objects by computer graphics, as a means to determine the brightness of the drawing point, an environment cube and object space for environment mapping surrounding the transparent object are provided respectively, and this object space is divided into arbitrary sizes Then, the divided space is classified into three spaces each of which the planes of the transparent objects intersect, a complete object external space, and a complete object internal space, and a buffer having an address corresponding to the coordinate value of the divided space The buffer is provided with means for storing a flag for identifying the three spaces, and a buffer for storing polygon information constituting the object plane included in the space in the divided space where the object plane intersects. Each means is used as a pre-process for starting the drawing of the transparent object, as well as the polygon normal at the drawing point, From the line-of-sight vector and the refractive index, a refraction vector passing through the transparent object and a straight line expression starting from the drawing point are obtained, and a flag stored in the buffer is read with respect to the divided space intersecting with the straight line. When the condition is tested and the divided space includes the polygon surface, the stored polygons are sequentially read from the buffer and the intersection calculation with the straight line formula is performed.If there is no intersection, the straight line is further stepped, If there is an intersection, a surface normal of the intersected point is obtained from the polygon information, and from this surface normal and the refraction vector, a means for obtaining an output line-of-sight vector from which the refraction vector exits from the object again, and this output line-of-sight vector Finds a point that intersects the surface of the environmental cube, and the luminance on the environmental cube surface is the luminance of the drawing point, respectively Transparent object display circuit consisting of means.   2. The circuit according to claim 1, wherein as a means for selecting a divided space through which a linear expression starting from a drawing point passes, a straight line for each coordinate is based on a coordinate axis (hereinafter referred to as a major axis) having the largest inclination of the linear expression. When the equation is obtained, the unit length of the divided space is substituted and stepped, the flag stored in the buffer corresponding to the divided space including the obtained coordinate point is read and tested, and the case inside and outside the complete object If there is a flag that intersects the object plane while stepping on the straight line expression, means for reading the polygon information from the buffer and calculating the intersection of the straight line expression and the surface polygon, and the polygon information includes , Vertex coordinate value, vertex surface normal, vertex texture coordinate value, means including transmittance and attenuation factor, read vertex coordinate value, and intersection coordinate value obtained by intersection calculation Al, determined the respective polygon information in the intersection, the relative luminance obtained in the environment cube surface, using said polygon information, transparent object display circuit having both respective means for determining the final brightness of the drawing point.   3. The circuit according to claim 1, wherein in the refraction calculation using the refractive index of each intersection including the drawing point, when the incident line of sight, the refraction and the output line of sight vector are totally reflected, the reflection vector is obtained and the total reflection point is obtained. After obtaining a straight line expression starting from, the flag of the divided space through which the straight line expression passes is tested by the same means as in claims 1 and 2, and if the object plane intersects, the intersection point is determined. Means for repeating and obtaining the refraction or reflection vector from that point again up to a predetermined number of times, and the intersection processing with the divided space when the straight line by the refraction or output line-of-sight vector reaches the outer divided space of the object space And a transparent object display circuit having means for terminating the calculation and moving to the intersection calculation with the environmental cube surface.   A computer graphic image apparatus using the transparent object display circuit according to claim 1.

Images