Sunteți pe pagina 1din 2

1 Modeling primitives

1.1 3D space

1.1.1 2D versus 3D
Human beings live in a 3D world, but when they draw they work with a 2D
representation of the 3D world. There are two approaches to this:

1. Represent only planar faces of an object, e.g., the face of a house


2. Attempt to draw the scene using perspective laws

Since in computer graphics the graphics devices (screens) are 2D, both of these
methods are used, each corresponding to a different approach to modeling, graphic
systems, and applications.
A graphics system is said to be 2D if the internal representation of the graphic
information in the computer is 2D; a graphics system is 3D when the computer has
knowledge of 3D information. There is a fundamental difference between the two
approaches. When an image of a house in perspective is displayed on a screen, it may
not be possible to determine whether the image was produced with a 2D or a 3D
system. The house may have been drawn in perspective and the projection entered
into the computer using a 2D graphics system; in this case, the role of the system is
simply to display the image. The house may also have been generated by a 3D
graphics system using 3D data; in this case, the system calculated the projections
before displaying the image. It is important to note that when we discuss 3D images
in this book, we are referring to images based on 3D models stored by the computer
and not 3D techniques like stereoscopic views (see Sect. 2.4.4) or holography.
It should be noted that 2D space can be viewed as a special case of 3D space, with
the third dimension Z always zero. For this reason, we shall use a 3D system in such
a way that the third dimension is simply added to the 2D system as shown in Fig.
1.1.

1.1.2 Modeling primitives, external and internal models


Typically, object representations may be classified into three levels:

- Wire-frame representations
- Surface representations
- Volume representations

N. Magnenat-Thalmann et al., Image Synthesis


© Springer-Verlag Tokyo 1987
2 Modeling primitives

y
YA~

4 4 P= <4,
P=<3.8, 2.8>
3 3

2·- 2

1 f-
..
,...
1

4
0 1 2 3 4 X 0 1 2 3 X
a b
Fig. 1.1. Coordinate system. a2D; b 3D

In a wire-frame representation (see Sect. 1.2.1), objects are merely


represented by a set of segment lines, typically by vertices and edges, with no data on
the surfaces and volumes. Such a representation is frequently too simple and cannot
be used to achieve a high degree of realism.
In a surface representation, objects are represented by a set of primitive
surfaces. Typically, three categories of description are used, depending on the object
shape:
- Description by a set of polygons (see Sect. 1.2.2)
- Description by the equation of an algebraic surface (see Sect. 1.2.3 and Chap. 3)
- Description by patches (see Chap. 3)
In a solid representation (see Sect. 1.2.5 and Chap. 4), objects are essentially
considered as a set of primitive volumes or as a bounded portion of the 3D space.
When a user builds an object, certain things have to be known about the model
used to represent the object. This model, which requires some knowledge about the
object, is called an external model. A high level external model is a procedural
model, which is in fact a model based on data abstraction (see Sect. 1.4). In a
procedural model, the object is not defined explicitly but is represented by a piece of
code (a procedure or a high-level graphic type) to which parameters are passed. For
example, a procedural model for a box (parallelepiped) may only require four
vertices, as shown in Fig. 1.2.
Then, the object is encoded according to an internal model. In our example,
the cube may be defmed:
- As a wire-frame: 12 segment lines (edges) defined by vertex coordinates
- As a surface: six polygonal facets defmed by vertex coordinates
- As a volume: a solid defined by its boundary representation, a space subdivision,
or as a volume primitive
Moreover, an object may be converted from a geometric internal model to a
display internal model. For example, a parametric surface may be defined by a
few coefficients, but converted into a polygonal planar facet internal model for the
display step.
Image-synthesis systems. often convert all external models into a unique internal
representation, because this facilitates the manipulation and transformation of objects.

S-ar putea să vă placă și