JPH08329280A - Image synthesizer - Google Patents

Abstract

(57) [Summary] [Purpose] To synthesize a high-quality two-dimensional image with good reproducibility of the edge information of the object. A model information storage unit 1 that stores model information representing a three-dimensional structure of an object, a rendering information generation unit 2 that outputs rendering information for converting the model information into a two-dimensional image, and the rendering information. It is configured by a polygon data generation unit 10 that inputs model information and generates polygon data for synthesizing a two-dimensional image, and an image synthesizing unit 3 that inputs the polygon data and synthesizes a two-dimensional image. The unit 10 generates at least primary polygon data from the model information, extracts some polygons from each of the primary polygons according to the rendering information, and subdivides the polygons.
Generate the next polygon.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image synthesizing apparatus for synthesizing a two-dimensional image based on the three-dimensional structure of an object.

[0002]

2. Description of the Related Art A so-called three-dimensional computer graphics (CG) technique for synthesizing a two-dimensional image from a three-dimensional structure of an object using a computer is very important for data visualization in CAD / CAM and the like. It is a technology. Conventionally, in the technique of three-dimensional CG, the amount of calculation required for synthesizing a secondary image differs depending on the type of three-dimensional model information of one object, and is generally required for synthesizing a secondary image of high image quality. The amount of calculation tends to increase. Specifically, when a plurality of polygon information is generated from a three-dimensional structure and a two-dimensional image is combined, a high-quality secondary image can be combined if the number of polygons is large and the number of polygons is small. There was a problem that the amount of calculation required to synthesize the next image was large.

As an improvement technique for this, for example, there is a technique proposed in 1993 Japanese Patent Application No. 105689.
This technique will be described with reference to FIGS. 9 and 10. This is because in the model information storage unit 91 shown in FIG. 9, three-dimensional model information 110 represented in advance by a rough polygon for one object and model information 111 represented by a fine polygon in advance as shown in FIG. It is stored as three-dimensional coordinate data of each vertex. The rendering information generation unit 92 generates information other than the model information of the object, for example, the position and direction of the object from the projection surface, and surrounding illumination information. In this example, based on this rendering information, a plurality of model information items are selected and switched by the selection unit 93 as needed, so that the amount of calculation is reduced.

[0004]

However, the conventional technique has a problem that a large amount of calculation is still required to obtain a secondary image with high image quality for one object. For example, by simply increasing the number of polygons from the rough three-dimensional model information 110 shown in FIG. 10 to a double-precision model 111 in one direction,
When synthesizing an image that is twice as precise, the number of polygons must be quadrupled, so the amount of calculation required is also quadrupled, and thus it takes a lot of time to synthesize a high-quality image. .

An object of the present invention is to solve the above problems, and an object thereof is to provide an image synthesizing apparatus capable of synthesizing a high quality two-dimensional image with good reproducibility of edge information of an object.

[0006]

In order to achieve the above object, the present invention has a model information storage means for storing model information representing a three-dimensional structure of an object, and a rendering for converting the model information into a two-dimensional image. A rendering information generating unit that outputs information, a polygon data generating unit that inputs the rendering information and the model information, generates polygon data for synthesizing a two-dimensional image, and the polygon data as an input, a two-dimensional image And an image synthesizing unit for synthesizing the polygon data, wherein the polygon data generating unit generates at least a primary polygon data from the model information, and a part of each of the primary polygons according to the rendering information. This is a configuration having means for generating a secondary polygon by extracting a polygon and subdividing it.

[0007]

The most important factor that determines the image quality of a two-dimensional image is the reproducibility of the edge information of the object. By generating fine polygon data from the three-dimensional model information, high-quality two-dimensional image synthesis can be obtained because the reproducibility of the edge information of the object is improved.

According to the present invention, from among the primary polygon data,
By extracting the object contour forming polygons, which are important for reproducing the edge information of this object, and generating the subdivided secondary polygons only from this object contour forming polygon, the overall number of polygons is overwhelmingly reduced. It is possible to synthesize a high-quality two-dimensional image with good reproducibility of the edge information of the object. As a result, the image quality is very high with a very small calculation amount.
It is possible to synthesize three-dimensional images.

[0009]

1 is a block diagram of an image synthesizing apparatus showing a first embodiment of the present invention. In FIG. 1, 1 is a model information storage unit, 2 is a rendering information generation unit, 3 is an image synthesis unit, 4 is a primary polygon generation unit, 5 is an object contour formation polygon extraction unit, 6 is a secondary polygon generation unit, and 8 Is a polygon data selection unit, and 10 is a polygon data generation unit.

FIG. 2 is a schematic diagram showing the operation of the image processing apparatus showing the first embodiment of the present invention. The operation will be described below with reference to FIGS. 1 and 2. First, in the model information storage unit 1, as shown in the range data 11 and the texture data 12 in FIGS. 2A and 2B, the model information of the object is
The range data 11 in the form of the radius data r taken in the cylindrical coordinate system X-Φ as shown in (c) and the texture information 12 of the object surface at the position corresponding thereto are stored.

In the rendering information generator 2, information other than the model information of the object, for example, the position of the object from the projection plane,
Direction, ambient lighting information, and the like are generated. The polygon data generation unit 10 generates the primary polygon data 14 composed of each vertex shown on X-Φ from the model information. At this time, the information obtained from the model information is: polygon P: each vertex {p0, p1, p2} vertex pn (n = 0, 1, 2): {three-dimensional coordinate (X, Y, Z)}: {texture information 12 is the coordinates (u, v)}, but from this {three-dimensional coordinates (X, Y, Z)} information, as shown in FIG. x, y)} is calculated. This primary polygon data 1
From 4, it is possible to synthesize a two-dimensional image of a rough object as shown at 15 in FIG.

In the object contour forming polygon extraction unit 5, first, the relationship between the normal vector of each primary polygon and the normal vector of the two-dimensional image projection plane is examined as shown in FIG. The primary polygon having a negative value is identified as the back polygon 101, and the primary polygon having a positive value is identified as the front polygon 102. Furthermore, the table polygon 10
Among the primary polygons identified as 2, the polygon whose adjacent polygon is the back polygon 101 is extracted as the object contour forming polygon, and the line shared with the back polygon is extracted as the object contour line 103.

In the secondary polygon generator 6, at least one or more secondary polygon points 104 are provided on the object contour line 103 in the object contour forming polygon, as shown in FIG. 2 (g). 105 is generated. At this time, the three-dimensional position of the point 104 on the object contour line 103 is generated from the range data 11.

The polygon data generation unit 10 uses this secondary polygon and one of the surface polygons other than the object contour forming polygon.
Outputs the rendering information generated from the next polygon. The output format is, for each polygon, polygon P: each vertex {p0, p1, p2} vertex pn (n = 0, 1, 2): {XY coordinates (x, y) on the projection plane} {on texture information 12 Coordinates (u, v)} at.

In the image synthesizing unit 3, according to the polygon data input from the polygon data generating unit 10, the texture data 1 of the texture information is displayed on the XY coordinates on the projection plane.
The image of the object is composed by transferring 2.

As shown in FIG. 2 (h), in the image of the object synthesized by this, the object contour portion viewed from the projection plane is reproduced very finely by the secondary polygon 105, and FIG. Compared with the two-dimensional synthetic image 15 of the rough object synthesized only from the primary polygon data shown in (1), a high-quality two-dimensional synthetic image 18 can be synthesized in each stage. Moreover, the number of polygons must be quadrupled in order to double the precision of the contour portion by simply increasing the number of polygons. However, according to the present invention, the number of polygons increased by this is equal to that of the contour portion. In order to double the precision of the contour portion, it is only necessary to double the polygon of the contour portion. Therefore, the increase in the amount of calculation caused by the improvement in image quality can be suppressed to a minimum.

When the object position is far from the image projection surface, the size of the image becomes small, or when the object moves fast on the screen, it is not necessary to accurately reproduce the contour information visually. There is. In this case, the polygon data selection unit 8 shown in FIG. 1 selects a primary polygon to increase the speed.

Here, the model information storage unit 1 stores range data 11 in the form of radius data taken in a cylindrical coordinate system and texture data 12 of the object surface at a position corresponding to the model information of the object. However, instead of the range data 11, as shown in FIG. 2 (i), the vertex coordinates of the primary polygon and the vertex coordinates on each line of the primary polygon that can be the vertices of the secondary polygon. It may be in the form of memorizing. In this case, the stored model information format is as follows for each primary polygon: Is. Here, the difference between the vertex data of FIG. 2 and the data shown in the conventional fine polygon 111 of FIG.
The secondary polygon vertices are located on the line of the primary polygon, and the secondary polygon is divided and generated for the first time depending on the direction and position where the image is combined with the projection surface.

FIG. 3 is a block diagram of an image synthesizing apparatus showing a second embodiment of the present invention. In FIG. 3, 31 is a model information storage unit, 32 is a rendering information generation unit, 33 is an image synthesis unit, 34 is a primary polygon generation unit, 35 is an object contour forming polygon extraction unit, 36 is a secondary polygon generation unit, and 3
Reference numeral 7 is a polygon data generator.

FIG. 4 is a schematic diagram showing the operation of the image processing apparatus in the second embodiment of the present invention. 3 and 4 below
The operation will be described using. First, in the model information storage unit 31, the model information of the object is range data 1 in the form of radius data r taken in the cylindrical coordinate system X-Φ shown in FIG.
1 and the texture information 1 of the object surface at the corresponding position 1
2 is stored. The rendering information generating unit 32 generates information other than the model information of the object, for example, the position and direction of the object from the projection surface, and surrounding illumination information. Further, the polygon data generation unit 37 generates the primary polygon data 14 from the model information as shown on X-Φ in FIG. From this primary polygon data 14, a two-dimensional synthetic image 15 of a rough object can be synthesized as shown in FIG.

The object contour forming polygon extraction unit 35 first investigates the relationship between the normal vector of each primary polygon and the normal vector of the two-dimensional image projection plane as shown in FIG. The primary polygon having a negative value is identified as the back polygon 101, and the primary polygon having a positive value is identified as the front polygon 102. Further, among the primary polygons identified as the table polygons, those whose inner product of both vectors has a value of 0 are extracted as the polygons 120 near the object contour, and each line of the polygons 120 near the object contour is defined as the object contour line 121. Extract.

As shown in FIG. 4G, the secondary polygon generator 36 first divides the object contour neighborhood polygon 120, and the table polygon 122 adjacent to the object contour neighborhood polygon 120 is also on the object contour line 121. A secondary polygon is generated by providing at least one or more points in. At this time, the three-dimensional position of the point on the object contour line is generated from the range data 11. The polygon data generator 37 outputs polygon data generated from the secondary polygon and the primary polygon.

In accordance with the polygon data input, the image synthesizing section 33 synthesizes an image of the object by transferring the data of the texture information 12 onto the XY coordinates on the projection surface.

As shown in FIG. 4 (h), in the image of the synthesized object, the contour portion of the object seen from the projection plane is reproduced very finely by the secondary polygon 123. Compared with the two-dimensional image of the rough object synthesized only from the primary polygon data 14 shown in (1), a high-quality two-dimensional image can be synthesized at each stage.

FIG. 5 is a block diagram of an image synthesizing apparatus showing a third embodiment of the present invention. In FIG. 5, 51 is a model information storage unit, 52 is a rendering information generation unit, 53 is an image synthesis unit, 54 is a primary polygon generation unit, 55 is an object contour forming line extraction unit, 56 is a secondary polygon generation unit, and 57.
Is a polygon data generator.

FIG. 6 is a schematic diagram showing the operation of the image processing apparatus showing the third embodiment of the present invention. The operation will be described below with reference to FIGS. First, the model information storage unit 5
In FIG. 1, model information of an object as shown in FIGS. 6 (a) and 6 (b) is shown as range data 11 in the form of radius data r in the cylindrical coordinate system X-Φ as shown in FIG. 6 (c). About Fig. 6 (i)
As shown in, the vertex coordinates of the primary polygon and the vertex coordinates on each line of the primary polygon that can be the vertex of the secondary polygon are stored. In this case, the model information format stored is: for each primary polygon, polygon P: each vertex {p0, p1, p2} each line {l0, l1, l2} vertex pn (n = 0, 1, 2): {XYZ coordinates (x, y, z)} {Coordinates on texture information 12 (u, v) Line ln (n = 0, 1, 2): {Adjacent polygon P '} {Secondary polygon vertex p' m (m = 0 to)} Vertex for secondary polygon p'm (m = 0 to): {XYZ coordinates (x, y, z)} {Position ratio on primary polygon line: 0 to 1.0} {Shift of XYZ coordinates from the primary polygon line Vd = (dx, dy, dz)}.

However, the {coordinate (u, v) on the texture information 12} of the vertex for the secondary polygon is obtained by interpolation from the position ratio on the primary polygon line and the vertex information of the primary polygon.

In the object contour forming line extraction unit 55, first, for each line of each primary polygon, it is obtained from the deviation vector Vd of the XYZ coordinates from the primary polygon line of the secondary polygon vertex p'm and the rendering information. 2
From the relation between the normal vector V of the plane of the three-dimensional image projection and the distance Z,
The magnitude rd of the positional deviation due to Vd on the projection surface is calculated by the following equation.

Rd = | V × Vd | / Z If the value of rd exceeds a certain threshold value, the secondary vertex 13
Extract as 1. In the secondary polygon generator 56, as shown in FIG.
As shown in (g), first, for each line of each primary polygon, the secondary vertex 131 is searched, and when the secondary vertex 131 exists, the secondary polygon 131 is used to divide the primary polygon. As a result, the secondary polygon 132 is generated.

The polygon data generator 57 outputs the polygon data generated from the secondary polygon and the primary polygon, and the image synthesizer 53 inputs the polygon data to the XY coordinates on the projection plane. The image of the object is synthesized by transferring the data of the texture information 12.

As shown in FIG. 6 (h), the image of the object synthesized in this way is such that the object contour portion as viewed from the projection surface is reproduced very finely by the secondary polygon 132, and only the primary polygon data is obtained. Compared with the two-dimensional image of a rough object synthesized from, a high-quality two-dimensional image can be synthesized at each stage.

FIG. 7 is a block diagram of an image synthesizing apparatus showing a fourth embodiment of the present invention. In FIG. 7, 71 is a model information storage unit, 72 is a rendering information generation unit, 73 is an image synthesis unit, 74 is a primary polygon generation unit, 75 is an object contour forming polygon extraction unit, 76 is a secondary polygon generation unit, 7
Reference numeral 7 is a tertiary polygon generation unit, 78 is a selection unit, and 79 is a polygon data generation unit.

FIG. 8 is a schematic diagram showing the operation of the image processing apparatus according to this embodiment. In this embodiment, the operation up to the secondary polygon generator 76 is the same as in the first embodiment,
From the output secondary polygon data, as shown in FIG. 8A, a considerably precise contour portion of the object can be synthesized. However, in the contour portion in the composite image, each line of the secondary polygon 105 causes a corner to be perceived at some places, which causes visual discomfort.

In the cubic polygon generator 77, as shown in FIG. 8B, the object contour line of each line of the secondary polygon 105 is extracted as a column of its vertices 108, and interpolation points are included in this column. By providing 107, the object contour on the projection surface is smoothed. As a result, a part of the secondary polygon data 105 is further subdivided, and the tertiary polygon 106 is generated.

As a method of providing the interpolation point 107, several methods using a spline function or the like have been proposed. Any of them may be used here. However, if the distance from the line before division of the interpolation point 107 does not exceed a certain value, the interpolation point is canceled and the line is not divided.

As shown in FIG. 8 (c), the image of the object synthesized by this not only reproduces the object contour portion when viewed from the projection plane very finely by the secondary polygon 105, but also The contour portion is smoothly reproduced, and the corner of the contour portion peculiar to the image synthesized from the polygon data is hardly perceived, so that a more visually natural two-dimensional image can be synthesized.

Moreover, since the number of polygons is obtained by only subdividing a part of the contour portion, it is possible to minimize the increase in the amount of calculation caused by the improvement in image quality. When the object is an artificial object such as a dice, such smoothing may be adversely affected and the reproducibility may be impaired. This effect is prevented by providing an On / Off fragment for smoothing as one of the attributes for each line of the primary Boligon as the three-dimensional model information.

In this case, the model information format stored is
For each primary polygon, polygon P: each vertex {p0, p1, p2} each line {l0, l1, l2} vertex pn (n = 0, 1, 2): {XYZ coordinates (x, y, z)} {Coordinates on texture information 12 (u, v) Line ln (n = 0, 1, 2): {Adjacent polygon P '} {Secondary polygon vertex p'm (m = 0)} {Smoothing fragment (On / Off)} Vertex for secondary polygon p'm (m = 0 to): {XYZ coordinates (x, y, z)} {Position ratio on primary polygon line: 0 to 1.0} {1 Deviation of XYZ coordinates from the next polygon line Vd = (dx, dy, dz)}.

[0039]

As described above, according to the present invention, the edge information of the most important object that determines the image quality of a two-dimensional image can be accurately reproduced, and it is important to reproduce the edge information of the object. Polygons important for contour formation are extracted, and only the polygons important for contour formation of this object are further subdivided 2
By generating the next polygon, it is possible to synthesize a high-quality two-dimensional image with good reproducibility of the edge information of the object by using an overwhelmingly small number of polygons.

Moreover, in order to make the contour portion twice as precise by simply increasing the number of polygons, the number of polygons is set to four.
According to the present invention, the number of polygons increased by this is a part of the contour portion, and in order to double the precision of the contour portion, the polygon of the contour portion is only doubled. Therefore, it is possible to minimize the increase in the amount of calculation that occurs due to this improvement in image quality.

Further, not only the object contour portion when viewed from the projection surface can be accurately reproduced, but also the object contour portion can be smoothly reproduced, which is a cause of deterioration peculiar to an image synthesized from polygon data. Since it is difficult for the corners of the contour to be perceived, it is possible to synthesize a visually more natural two-dimensional image.

Thus, according to the present invention, CA
When synthesizing images in D, CAM, games, etc., it becomes possible to synthesize extremely high-quality images at very high speed with an inexpensive device or computer, and the industrial effect thereof is great.

[Brief description of drawings]

FIG. 1 is a block diagram of an image synthesizing apparatus showing a first embodiment of the present invention.

2A to 2I are schematic diagrams showing the operation of the image processing apparatus showing the first embodiment of the present invention.

FIG. 3 is a block diagram of an image synthesizing apparatus showing a second embodiment of the present invention.

4A to 4H are schematic diagrams showing an operation of an image processing apparatus showing a second embodiment of the present invention.

FIG. 5 is a block diagram of an image synthesizing device showing a third embodiment of the present invention.

6A to 6I are schematic diagrams showing the operation of the image processing apparatus showing the third embodiment of the present invention.

FIG. 7 is a block diagram of an image synthesizing apparatus showing a fourth embodiment of the present invention.

8A to 8C are schematic diagrams showing the operation of an image processing apparatus showing a fourth embodiment of the present invention.

FIG. 9 is a block diagram of an image synthesizing device showing a conventional example.

10A to 10D are schematic diagrams showing the operation of an image processing apparatus showing a conventional example.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Model information storage unit 2 Rendering information generation unit 3 Image synthesis unit 4 Primary polygon generation unit 5 Object contour formation polygon extraction unit 6 Secondary polygon generation unit 8 Polygon data selection unit 10 Polygon data generation unit 11 Range data 12 Texture data 14 Primary polygon data 15 Composite image by primary polygon 18 Composite image 55 Contour forming line extraction unit 77 Third polygon generating unit 101 Back polygon 102 Front polygon 103 Object contour line 104 Secondary polygon vertex 105 Secondary polygon 106 Tertiary polygon 106

Claims (9)

[Claims] 1. Model information storage means for storing model information representing a three-dimensional structure of an object, rendering information generation means for outputting rendering information for converting the model information into a two-dimensional image, and the rendering information. Polygon data generating means for generating polygon data for synthesizing a two-dimensional image by inputting model information and image synthesizing means for synthesizing a two-dimensional image by inputting the polygon data are provided. The generating means is a means for generating at least primary polygon data from the model information, and a secondary polygon by extracting some polygons from each of the primary polygons according to the rendering information and subdividing the polygons. An image synthesizing device having means for generating 2. Polygon data generating means, means for generating primary polygon data from model information, projection of a two-dimensional image to be synthesized, which is obtained from at least the normal vector of this primary polygon data and the rendering information. Based on the condition of the relation between the surface normal vectors, each primary polygon is sorted into a back polygon and a front polygon in the object, and when adjacent polygons are back polygons in each front polygon, this is extracted as an object contour forming polygon. The image synthesizing apparatus according to claim 1, further comprising means and means for generating a secondary polygon obtained by further subdividing a primary polygon with respect to the object contour forming polygon. 3. The polygon data generating means is at least
Means for generating primary polygon data from model information,
At least this normal vector of the primary polygon data,
Means for extracting a polygon near the object contour, and further subdividing the primary polygon for the polygon near the object contour, on condition of the relation of the normal vector of the projection plane of the two-dimensional image to be synthesized, which is obtained from the rendering information. Done 2
The image synthesizing apparatus according to claim 1, further comprising means for generating a next polygon. 4. The image synthesizing apparatus according to claim 1, wherein the model information has at least primary polygon data and vertex data for secondary polygons on the primary polygons as three-dimensional information. 5. The model information is three-dimensional information, and at least primary polygon data and two lines on each line of the primary polygon are used.
The image synthesizing apparatus according to claim 1, further comprising vertex data for the next polygon. 6. The model information includes, as three-dimensional information, primary polygon data and vertex data for secondary polygons on the primary polygons, and the polygon data generation unit uses at least the primary polygon data based on the model information. And the deviation of the position of the vertices for the secondary polygon from the primary polygon plane exceeds a certain threshold size projected on the projection plane of the two-dimensional image to be synthesized obtained from the rendering information. 6. The image synthesizing apparatus according to claim 4, further comprising means for detecting a vertex and generating a secondary polygon obtained by subdividing the primary polygon using the vertex for the secondary polygon. 7. The polygon data generating means includes at least means for generating primary polygon data from model information, and a case where the primary polygon data is projected on a projection surface of a two-dimensional image to be combined obtained from rendering information. Object contour forming line extracting means for extracting a line forming an object contour from each line of each primary polygon data as an object contour forming line, and object contour forming by providing interpolation points on the object contour forming line The image synthesizing apparatus according to claim 1, further comprising secondary polygon generation means for generating a secondary polygon obtained by subdividing a primary polygon having a line. 8. The object contour forming line extracting means sets the back surface of the object for each primary polygon based on the condition of the relationship between the normal vector of the primary polygon data and the normal vector of the projection plane of the two-dimensional image to be synthesized. The feature is that it is sorted into polygons and front polygons, and when adjacent polygons of each front polygon are back polygons, this is extracted as an object contour forming polygon, and a line shared with the back polygon is extracted as an object contour forming line. The image synthesizing apparatus according to claim 7. 9. The image synthesizing apparatus according to claim 1, wherein the model information has three-dimensional information including at least polygon data and fragment data which can be divided or cannot be divided as attribute values of each line of each polygon.