1 | \chapter{Introduction}%\chapter
|
---|
2 |
|
---|
3 | \label{chap:introduction}
|
---|
4 |
|
---|
5 | \section{Structure of the report}
|
---|
6 |
|
---|
7 | The report consists of two introductory chapters, which provide a
|
---|
8 | theoretical background for description of the algorithms, and three
|
---|
9 | chapters dealing with the actual visibility algorithms.
|
---|
10 |
|
---|
11 | This chapter provides an introduction to visibility by using a
|
---|
12 | taxonomy of visibility problems and algorithms. The taxonomy is used
|
---|
13 | to classify the later described visibility
|
---|
14 | algorithms. Chapter~\ref{chap:analysis} provides an analysis of
|
---|
15 | visibility in 2D and 3D polygonal scenes. This analysis also includes
|
---|
16 | formal description of visibility using \plucker coordinates of
|
---|
17 | lines. \plucker coordinates are exploited later in algorithms for
|
---|
18 | mutual visibility verification (Chapter~\ref{chap:mutual}).
|
---|
19 |
|
---|
20 | Chapter~\ref{chap:online} describes a visibility culling algorithm
|
---|
21 | used to implement the online visibility culling module. This
|
---|
22 | algorithm can be used accelerate rendering of fully dynamic scenes
|
---|
23 | using recent graphics hardware. Chapter~\ref{chap:sampling}
|
---|
24 | describes global visibility sampling algorithm which forms a core of
|
---|
25 | the PVS computation module. This chapter also describes view space
|
---|
26 | partitioning algorithms used in close relation with the PVS
|
---|
27 | computation. Finally, Chapter~\ref{chap:mutual} describes mutual
|
---|
28 | visibility verification algorithms, which are used by the PVS
|
---|
29 | computation module to generate the final solution for precomputed
|
---|
30 | visibility.
|
---|
31 |
|
---|
32 |
|
---|
33 |
|
---|
34 |
|
---|
35 | %% summarize the state of the art
|
---|
36 | %% visibility algorithms and their relation to the proposed taxonomy of
|
---|
37 | %% visibility problems. The second part of the survey should serve as a
|
---|
38 | %% catalogue of visibility algorithms that is indexed by the proposed
|
---|
39 | %% taxonomy of visibility problems.
|
---|
40 |
|
---|
41 |
|
---|
42 |
|
---|
43 | \section{Domain of visibility problems}
|
---|
44 | \label{sec:prob_domain}
|
---|
45 |
|
---|
46 | Computer graphics deals with visibility problems in the context of
|
---|
47 | 2D, \m25d, or 3D scenes. The actual problem domain is given by
|
---|
48 | restricting the set of rays for which visibility should be
|
---|
49 | determined.
|
---|
50 |
|
---|
51 | Below we list common problem domains used and the corresponding domain
|
---|
52 | restrictions:
|
---|
53 |
|
---|
54 | \begin{enumerate}
|
---|
55 | \item
|
---|
56 | {\em visibility along a line}
|
---|
57 | \begin{enumerate}
|
---|
58 | \item line
|
---|
59 | \item ray (origin + direction)
|
---|
60 | \end{enumerate}
|
---|
61 | \newpage
|
---|
62 | \item
|
---|
63 | {\em visibility from a point} ({\em from-point visibility})
|
---|
64 | \begin{enumerate}
|
---|
65 | \item point
|
---|
66 | \item point + restricted set of rays
|
---|
67 | \begin{enumerate}
|
---|
68 | \item point + raster image (discrete form)
|
---|
69 | \item point + beam (continuous form)
|
---|
70 | \end{enumerate}
|
---|
71 | \end{enumerate}
|
---|
72 |
|
---|
73 |
|
---|
74 | \item
|
---|
75 | {\em visibility from a line segment} ({\em from-segment visibility})
|
---|
76 | \begin{enumerate}
|
---|
77 | \item line segment
|
---|
78 | \item line segment + restricted set of rays
|
---|
79 | \end{enumerate}
|
---|
80 |
|
---|
81 | \item
|
---|
82 | {\em visibility from a polygon} ({\em from-polygon visibility})
|
---|
83 | \begin{enumerate}
|
---|
84 | \item polygon
|
---|
85 | \item polygon + restricted set of rays
|
---|
86 | \end{enumerate}
|
---|
87 |
|
---|
88 | \item
|
---|
89 | {\em visibility from a region} ({\em from-region visibility})
|
---|
90 | \begin{enumerate}
|
---|
91 | \item region
|
---|
92 | \item region + restricted set of rays
|
---|
93 | \end{enumerate}
|
---|
94 |
|
---|
95 | \item
|
---|
96 | {\em global visibility}
|
---|
97 | \begin{enumerate}
|
---|
98 | \item no further input (all rays in the scene)
|
---|
99 | \item restricted set of rays
|
---|
100 | \end{enumerate}
|
---|
101 | \end{enumerate}
|
---|
102 |
|
---|
103 | The domain restrictions can be given independently of the dimension
|
---|
104 | of the scene, but the impact of the restrictions differs depending on
|
---|
105 | the scene dimension. For example, visibility from a polygon is
|
---|
106 | equivalent to visibility from a (polygonal) region in 2D, but not in
|
---|
107 | 3D.
|
---|
108 |
|
---|
109 | %*****************************************************************************
|
---|
110 |
|
---|
111 | \section{Dimension of the problem-relevant line set}
|
---|
112 |
|
---|
113 | The six domains of visibility problems stated in
|
---|
114 | Section~\ref{sec:prob_domain} can be characterized by the {\em
|
---|
115 | problem-relevant line set} denoted ${\cal L}_R$. We give a
|
---|
116 | classification of visibility problems according to the dimension of
|
---|
117 | the problem-relevant line set. We discuss why this classification is
|
---|
118 | important for understanding the nature of the given visibility problem
|
---|
119 | and for identifying its relation to other problems.
|
---|
120 |
|
---|
121 |
|
---|
122 | For the following discussion we assume that a line in {\em primal
|
---|
123 | space} can be mapped to a point in {\em line space}. For purposes of
|
---|
124 | the classification we define the line space as a vector space where a
|
---|
125 | point corresponds to a line in the primal space\footnote{A classical
|
---|
126 | mathematical definition says: Line space is a direct product of two
|
---|
127 | Hilbert spaces~\cite{Weisstein:1999:CCE}. However, this definition
|
---|
128 | differs from the common understanding of line space in computer
|
---|
129 | graphics~\cite{Durand99-phd}}.
|
---|
130 |
|
---|
131 |
|
---|
132 |
|
---|
133 | \subsection{Parametrization of lines in 2D}
|
---|
134 |
|
---|
135 | There are two independent parameters that specify a 2D line and thus
|
---|
136 | the corresponding set of lines is two-dimensional. There is a natural
|
---|
137 | duality between lines and points in 2D. For example a line expressed
|
---|
138 | as: $l:y=ax+c$ is dual to a point $p=(-c,a)$. This particular duality
|
---|
139 | cannot handle vertical lines. See Figure~\ref{fig:duality2d} for an
|
---|
140 | example of other dual mappings in the plane. To avoid the singularity
|
---|
141 | in the mapping, a line $l:ax+by+c=0$ can be represented as a point
|
---|
142 | $p_l=(a,b,c)$ in 2D projective space ${\cal
|
---|
143 | P}^2$~\cite{Stolfi:1991:OPG}. Multiplying $p_l$ by a non-zero scalar
|
---|
144 | we obtain a vector that represents the same line $l$. More details
|
---|
145 | about this singularity-free mapping will be discussed in
|
---|
146 | Chapter~\ref{chap:analysis}.
|
---|
147 |
|
---|
148 |
|
---|
149 | \begin{figure}[!htb]
|
---|
150 | \centerline{
|
---|
151 | \includegraphics[width=0.9\textwidth,draft=\DRAFTFIGS]{figs/duality2d}}
|
---|
152 | \caption{Duality between points and lines in 2D.}
|
---|
153 | \label{fig:duality2d}
|
---|
154 | \end{figure}
|
---|
155 |
|
---|
156 |
|
---|
157 |
|
---|
158 |
|
---|
159 |
|
---|
160 | To sum up: In 2D there are two degrees of freedom in description of a
|
---|
161 | line and the corresponding line space is two-dimensional. The
|
---|
162 | problem-relevant line set ${\cal L}_R$ then forms a $k$-dimensional
|
---|
163 | subset of ${\cal P}^2$, where $0\leq k \leq 2$. An illustration of the
|
---|
164 | concept of the problem-relevant line set is depicted in
|
---|
165 | Figure~\ref{fig:classes}.
|
---|
166 |
|
---|
167 |
|
---|
168 | \begin{figure}[htb]
|
---|
169 | \centerline{
|
---|
170 | \includegraphics[width=0.8\textwidth,draft=\DRAFTFIGS]{figs/classes}}
|
---|
171 | \caption{The problem-relevant set of lines in 2D. The ${\cal L}_R$ for
|
---|
172 | visibility along a line is formed by a single point that is a mapping
|
---|
173 | of the given line. The ${\cal L}_R$ for visibility from a point $p$ is
|
---|
174 | formed by points lying on a line. This line is a dual mapping of the
|
---|
175 | point $p$. ${\cal L}_R$ for visibility from a line segment is formed
|
---|
176 | by a 2D region bounded by dual mappings of endpoints of the given
|
---|
177 | segment.}
|
---|
178 | \label{fig:classes}
|
---|
179 | \end{figure}
|
---|
180 |
|
---|
181 |
|
---|
182 | \subsection{Parametrization of lines in 3D}
|
---|
183 |
|
---|
184 |
|
---|
185 | Lines in 3D form a four-parametric space~\cite{p-rsls-97}. A line
|
---|
186 | intersecting a given scene can be described by two points on a sphere
|
---|
187 | enclosing the scene. Since the surface of the sphere is a two
|
---|
188 | parametric space, we need four parameters to describe the line.
|
---|
189 |
|
---|
190 | The {\em two plane parametrization} of 3D lines describes a line by
|
---|
191 | points of intersection with the given two
|
---|
192 | planes~\cite{Gu:1997:PGT}. This parametrization exhibits a singularity
|
---|
193 | since it cannot describe lines parallel to these planes. See
|
---|
194 | Figure~\ref{fig:3dlines} for illustrations of the spherical and the
|
---|
195 | two plane parameterizations.
|
---|
196 |
|
---|
197 |
|
---|
198 | \begin{figure}[htb]
|
---|
199 | \centerline{
|
---|
200 | \includegraphics[width=0.78\textwidth,draft=\DRAFTFIGS]{figs/3dlines}}
|
---|
201 | \caption{Parametrization of lines in 3D. (left) A line can be
|
---|
202 | described by two points on a sphere enclosing the scene. (right) The
|
---|
203 | two plane parametrization describes a line by point of intersection
|
---|
204 | with two given planes.}
|
---|
205 | \label{fig:3dlines}
|
---|
206 | \end{figure}
|
---|
207 |
|
---|
208 | Another common parametrization of 3D lines are the {\em \plucker
|
---|
209 | coordinates}. \plucker coordinates of an oriented 3D line are a six
|
---|
210 | tuple that can be understood as a point in 5D oriented projective
|
---|
211 | space~\cite{Stolfi:1991:OPG}. There are six coordinates in \plucker
|
---|
212 | representation of a line although we know that the ${\cal L}_R$ is
|
---|
213 | four-dimensional. This can be explained as follows:
|
---|
214 |
|
---|
215 | \begin{itemize}
|
---|
216 | \item Firstly, \plucker coordinates are {\em homogeneous
|
---|
217 | coordinates} of a 5D point. By multiplication of the coordinates
|
---|
218 | by any positive scalar we get a mapping of the same line.
|
---|
219 | \item Secondly, only 4D subset of the 5D oriented projective space
|
---|
220 | corresponds to real lines. This subset is a 4D ruled quadric called
|
---|
221 | the {\em \plucker quadric} or the {\em Grassman
|
---|
222 | manifold}~\cite{Stolfi:1991:OPG,Pu98-DSGIV}.
|
---|
223 | \end{itemize}
|
---|
224 |
|
---|
225 | Although the \plucker coordinates need more coefficients they have no
|
---|
226 | singularity and preserve some linearities: lines intersecting a set of
|
---|
227 | lines in 3D correspond to an intersection of 5D hyperplanes. More
|
---|
228 | details on \plucker coordinates will be discussed in
|
---|
229 | Chapter~\ref{chap:analysis} and Chapter~\ref{chap:mutual} where they
|
---|
230 | are used to solve the from-region visibility problem.
|
---|
231 |
|
---|
232 | To sum up: In 3D there are four degrees of freedom in the
|
---|
233 | description of a line and thus the corresponding line space is
|
---|
234 | four-dimensional. Fixing certain line parameters (e.g. direction) the
|
---|
235 | problem-relevant line set, denoted ${\cal L}_R$, forms a
|
---|
236 | $k$-dimensional subset of ${\cal P}^4$, where $0\leq k \leq 4$.
|
---|
237 |
|
---|
238 |
|
---|
239 | \subsection{Visibility along a line}
|
---|
240 |
|
---|
241 | The simplest visibility problems deal with visibility along a single
|
---|
242 | line. The problem-relevant line set is zero-dimensional, i.e. it is
|
---|
243 | fully specified by the given line. A typical example of a visibility
|
---|
244 | along a line problem is {\em ray shooting}.
|
---|
245 |
|
---|
246 | A similar problem to ray shooting is the {\em point-to-point}
|
---|
247 | visibility. The point-to-point visibility determines whether the
|
---|
248 | line segment between two points is occluded, i.e. it has an
|
---|
249 | intersection with an opaque object in the scene. Point-to-point
|
---|
250 | visibility provides a visibility classification (answer 1a), whereas
|
---|
251 | ray shooting determines a visible object (answer 2a) and/or a point of
|
---|
252 | intersection (answer 3a). Note that the {\em point-to-point}
|
---|
253 | visibility can be solved easily by means of ray shooting. Another
|
---|
254 | constructive visibility along a line problem is determining the {\em
|
---|
255 | maximal free line segments} on a given line. See Figure~\ref{fig:val}
|
---|
256 | for an illustration of typical visibility along a line problems.
|
---|
257 |
|
---|
258 | \begin{figure}[htb]
|
---|
259 | \centerline{
|
---|
260 | \includegraphics[width=0.85\textwidth,draft=\DRAFTFIGS]{figs/val}}
|
---|
261 | \caption{Visibility along a line. (left) Ray shooting. (center) Point-to-point visibility. (right) Maximal free line segments between two points.}
|
---|
262 | \label{fig:val}
|
---|
263 | \end{figure}
|
---|
264 |
|
---|
265 |
|
---|
266 | \subsection{Visibility from a point}
|
---|
267 |
|
---|
268 | Lines intersecting a point in 3D can be described by two
|
---|
269 | parameters. For example the lines can be expressed by an intersection
|
---|
270 | with a unit sphere centered at the given point. The most common
|
---|
271 | parametrization describes a line by a point of intersection with a
|
---|
272 | given viewport. Note that this parametrization accounts only for a
|
---|
273 | subset of lines that intersect the viewport (see Figure~\ref{fig:vfp}).
|
---|
274 |
|
---|
275 | \begin{figure}[htb]
|
---|
276 | \centerline{
|
---|
277 | \includegraphics[width=0.6\textwidth,draft=\DRAFTFIGS]{figs/vfp}}
|
---|
278 | \caption{Visibility from a point. Lines intersecting a point can be described by a
|
---|
279 | point of intersection with the given viewport.}
|
---|
280 | \label{fig:vfp}
|
---|
281 | \end{figure}
|
---|
282 |
|
---|
283 | In 3D the problem-relevant line set ${\cal L}_R$ is a 2D subset of
|
---|
284 | the 4D line space. In 2D the ${\cal L}_R$ is a 1D subset of the 2D
|
---|
285 | line space. The typical visibility from a point problem is the visible
|
---|
286 | surface determination. Due to its importance the visible surface
|
---|
287 | determination is covered by the majority of existing visibility
|
---|
288 | algorithms. Other visibility from a point problem is the construction
|
---|
289 | of the {\em visibility map} or the {\em point-to-region visibility}
|
---|
290 | that classifies a region as visible, invisible, or partially visible
|
---|
291 | with respect to the given point.
|
---|
292 |
|
---|
293 |
|
---|
294 | \subsection{Visibility from a line segment}
|
---|
295 |
|
---|
296 | Lines intersecting a line segment in 3D can be described by three
|
---|
297 | parameters. One parameter fixes the intersection of the line with the
|
---|
298 | segment the other two express the direction of the line. The
|
---|
299 | problem-relevant line set ${\cal L}_R$ is three-dimensional and it can
|
---|
300 | be understood as a 2D cross section of ${\cal L}_R$ swept according
|
---|
301 | to the translation on the given line segment (see
|
---|
302 | Figure~\ref{fig:vls}).
|
---|
303 |
|
---|
304 |
|
---|
305 | \begin{figure}[htb]
|
---|
306 | \centerline{
|
---|
307 | \includegraphics[width=0.8\textwidth,draft=\DRAFTFIGS]{figs/vls}}
|
---|
308 | \caption{Visibility from a line segment. (left) Line segment, a
|
---|
309 | spherical object $O$, and its projections $O^*_0$, $O^*_{0.5}$, $O^*_{1}$
|
---|
310 | with respect to the three points on the line segment. (right)
|
---|
311 | A possible parametrization of lines that stacks up 2D planes.
|
---|
312 | Each plane corresponds to mappings of lines intersecting a given
|
---|
313 | point on the line segment.}
|
---|
314 | \label{fig:vls}
|
---|
315 | \end{figure}
|
---|
316 |
|
---|
317 | In 2D lines intersecting a line segment form a two-dimensional
|
---|
318 | problem-relevant line set. Thus for the 2D case the ${\cal L}_R$ is a
|
---|
319 | two-dimensional subset of 2D line space.
|
---|
320 |
|
---|
321 |
|
---|
322 | \subsection{Visibility from a region}
|
---|
323 |
|
---|
324 | Visibility from a region (or from-region visibility) involves the
|
---|
325 | most general visibility problems. In 3D the ${\cal L}_R$ is a 4D
|
---|
326 | subset of the 4D line space. In 2D the ${\cal L}_R$ is a 2D subset of
|
---|
327 | the 2D line space. Consequently, in the presented classification
|
---|
328 | visibility from a region in 2D is equivalent to visibility from a line
|
---|
329 | segment in 2D.
|
---|
330 |
|
---|
331 | A typical visibility from a region problem is the problem of {\em
|
---|
332 | region-to-region} visibility that aims to determine if the two given
|
---|
333 | regions in the scene are visible, invisible, or partially visible (see
|
---|
334 | Figure~\ref{fig:vfr}). Another visibility from region problem is the
|
---|
335 | computation of a {\em potentially visible set} (PVS) with respect to a
|
---|
336 | given view cell. The PVS consists of a set of objects that are
|
---|
337 | potentially visible from any point inside the view cell. Further
|
---|
338 | visibility from a region problems include computing form factors
|
---|
339 | between two polygons, soft shadow algorithms or discontinuity meshing.
|
---|
340 |
|
---|
341 |
|
---|
342 | \begin{figure}[htb]
|
---|
343 | \centerline{
|
---|
344 | \includegraphics[width=0.6\textwidth,draft=\DRAFTFIGS]{figs/vfr}}
|
---|
345 | \caption{Visibility from a region --- an example of the region-to-region
|
---|
346 | visibility. Two regions and two occluders $A$, $B$
|
---|
347 | in a 2D scene. In line space the region-to-region visibility can be
|
---|
348 | solved by subtracting the sets of lines $A^*$ and $B^*$
|
---|
349 | intersecting objects $A$ and $B$ from the lines intersecting both
|
---|
350 | regions.}
|
---|
351 | \label{fig:vfr}
|
---|
352 | \end{figure}
|
---|
353 |
|
---|
354 |
|
---|
355 | \subsection{Global visibility}
|
---|
356 |
|
---|
357 | According to the classification the global visibility problems can be
|
---|
358 | seen as an extension of the from-region visibility problems. The
|
---|
359 | dimension of the problem-relevant line set is the same ($k=2$ for 2D
|
---|
360 | and $k=4$ for 3D scenes). Nevertheless, the global visibility problems
|
---|
361 | typically deal with much larger set of rays, i.e. all rays that
|
---|
362 | penetrate the scene. Additionally, there is no given set of reference
|
---|
363 | points from which visibility is studied and hence there is no given
|
---|
364 | priority ordering of objects along each particular line from ${\cal
|
---|
365 | L}_R$. Therefore an additional parameter must be used to describe
|
---|
366 | visibility (visible object) along each ray.
|
---|
367 |
|
---|
368 |
|
---|
369 | \subsection{Summary}
|
---|
370 |
|
---|
371 | The classification of visibility problems according to the dimension
|
---|
372 | of the problem-relevant line set is summarized in
|
---|
373 | Table~\ref{table:classification3D}. This classification provides
|
---|
374 | means for understanding how difficult it is to compute, describe, and
|
---|
375 | maintain visibility for a particular class of problems. For example a
|
---|
376 | data structure representing the visible or occluded parts of the scene
|
---|
377 | for the visibility from a point problem needs to partition a 2D ${\cal
|
---|
378 | L}_R$ into visible and occluded sets of lines. This observation
|
---|
379 | conforms with the traditional visible surface algorithms -- they
|
---|
380 | partition a 2D viewport into empty/nonempty regions and associate each
|
---|
381 | nonempty regions (pixels) with a visible object. In this case the
|
---|
382 | viewport represents the ${\cal L}_R$ as each point of the viewport
|
---|
383 | corresponds to a line through that point. To analytically describe
|
---|
384 | visibility from a region a subdivision of 4D ${\cal L}_R$ should be
|
---|
385 | performed. This is much more difficult than the 2D
|
---|
386 | subdivision. Moreover the description of visibility from a region
|
---|
387 | involves non-linear subdivisions of both primal space and line space
|
---|
388 | even for polygonal scenes~\cite{Teller:1992:CAA,Durand99-phd}.
|
---|
389 |
|
---|
390 | \begin{table*}[htb]
|
---|
391 | \begin{small}
|
---|
392 | \begin{center}
|
---|
393 | \begin{tabular}{|l|c|l|}
|
---|
394 | \hline
|
---|
395 | \multicolumn{3}{|c|}{2D} \\
|
---|
396 | \hline
|
---|
397 | \mc{domain} & $d({\cal L}_R)$ & \mc{problems} \\
|
---|
398 | \hline
|
---|
399 | \hline
|
---|
400 | \begin{tabular}{l}visibility along a line\end{tabular} & 0 & \begin{tabular}{l}ray shooting, point-to-point visibility\end{tabular}\\
|
---|
401 | \hline
|
---|
402 | \begin{tabular}{l}visibility from a point\end{tabular} & 1 & \begin{tabular}{l}view around a point, point-to-region visibility\end{tabular}\\
|
---|
403 | \hline
|
---|
404 | \begin{tabular}{l} visibility from a line segment \\ visibility from region \\ global visibility \end{tabular}
|
---|
405 | & 2 & \begin{tabular}{l} region-to-region visibility, PVS \end{tabular}\\
|
---|
406 | \hline
|
---|
407 | \hline
|
---|
408 | \multicolumn{3}{|c|}{3D} \\
|
---|
409 | \hline
|
---|
410 | \mc{domain} & $d({\cal L}_R)$ & \mc{problems} \\
|
---|
411 | \hline
|
---|
412 | \hline
|
---|
413 | \begin{tabular}{l}visibility along a line\end{tabular} & 0 & \begin{tabular}{l} ray shooting, point-to-point visibility \end{tabular}\\
|
---|
414 | \hline
|
---|
415 | \begin{tabular}{l}from point in a surface\end{tabular} & 1 & \begin{tabular}{l} see visibility from point in 2D \end{tabular}\\
|
---|
416 | \hline
|
---|
417 | \begin{tabular}{l}visibility from a point\end{tabular} & 2 & \begin{tabular}{l} visible (hidden) surfaces, point-to-region visibility,\\
|
---|
418 | visibility map, hard shadows
|
---|
419 | \end{tabular} \\
|
---|
420 | \hline
|
---|
421 | \begin{tabular}{l}visibility from a line segment\end{tabular} & 3 & \begin{tabular}{l} segment-to-region visibility (rare) \end{tabular}\\
|
---|
422 | \hline
|
---|
423 | \begin{tabular}{l}visibility from a region\\global visibility\end{tabular} & 4 & \begin{tabular}{l} region-region visibility, PVS, aspect graph,\\
|
---|
424 | soft shadows, discontinuity meshing
|
---|
425 | \end{tabular} \\
|
---|
426 |
|
---|
427 | \hline
|
---|
428 | \end{tabular}
|
---|
429 | \end{center}
|
---|
430 | \end{small}
|
---|
431 | \caption{Classification of visibility problems in 2D and 3D according
|
---|
432 | to the dimension of the problem-relevant line set.}
|
---|
433 | \label{table:classification3D}
|
---|
434 | \end{table*}
|
---|
435 |
|
---|
436 |
|
---|
437 |
|
---|
438 |
|
---|
439 |
|
---|
440 | \section{Classification of visibility algorithms}
|
---|
441 |
|
---|
442 |
|
---|
443 | The taxonomy of visibility problems groups similar visibility problems
|
---|
444 | in the same class. A visibility problem can be solved by means of
|
---|
445 | various visibility algorithms. A visibility algorithm poses further
|
---|
446 | restrictions on the input and output data. These restrictions can be
|
---|
447 | seen as a more precise definition of the visibility problem that is
|
---|
448 | solved by the algorithm.
|
---|
449 |
|
---|
450 | Above we classified visibility problems according to the problem
|
---|
451 | domain and the desired answers. In this section we provide a
|
---|
452 | classification of visibility algorithms according to other
|
---|
453 | important criteria characterizing a particular visibility algorithm.
|
---|
454 |
|
---|
455 |
|
---|
456 | \subsection{Scene restrictions}
|
---|
457 | \label{sec:scene_restrictions}
|
---|
458 |
|
---|
459 | Visibility algorithms can be classified according to the restrictions
|
---|
460 | they pose on the scene description. The type of the scene description
|
---|
461 | influences the difficulty of solving the given problem: it is simpler
|
---|
462 | to implement an algorithm computing a visibility map for scenes
|
---|
463 | consisting of triangles than for scenes with NURBS surfaces. We list
|
---|
464 | common restrictions on the scene primitives suitable for visibility
|
---|
465 | computations:
|
---|
466 |
|
---|
467 |
|
---|
468 | \begin{itemize}
|
---|
469 | \item
|
---|
470 | triangles, convex polygons, concave polygons,
|
---|
471 |
|
---|
472 | \item
|
---|
473 | volumetric data,
|
---|
474 |
|
---|
475 | \item
|
---|
476 | points,
|
---|
477 |
|
---|
478 | \item
|
---|
479 | general parametric, implicit, or procedural surfaces.
|
---|
480 |
|
---|
481 | \end{itemize}
|
---|
482 |
|
---|
483 | Some attributes of scenes objects further increase the complexity of the visibility computation:
|
---|
484 |
|
---|
485 | \begin{itemize}
|
---|
486 |
|
---|
487 | \item
|
---|
488 | transparent objects,
|
---|
489 |
|
---|
490 | \item
|
---|
491 | dynamic objects.
|
---|
492 |
|
---|
493 | \end{itemize}
|
---|
494 |
|
---|
495 | The majority of analytic visibility algorithms deals with static
|
---|
496 | polygonal scenes without transparency. The polygons are often
|
---|
497 | subdivided into triangles for easier manipulation and representation.
|
---|
498 |
|
---|
499 | \subsection{Accuracy}
|
---|
500 | \label{sec:accuracy}
|
---|
501 |
|
---|
502 | Visibility algorithms can be classified according to the accuracy of
|
---|
503 | the result as:
|
---|
504 |
|
---|
505 | \begin{itemize}
|
---|
506 | \item exact,
|
---|
507 | \item conservative,
|
---|
508 | \item aggressive,
|
---|
509 | \item approximate.
|
---|
510 | \end{itemize}
|
---|
511 |
|
---|
512 |
|
---|
513 | An exact algorithm provides an exact analytic result for the given
|
---|
514 | problem (in practice however this result is typically influenced by
|
---|
515 | the finite precision of the floating point arithmetics). A
|
---|
516 | conservative algorithm overestimates visibility, i.e. it never
|
---|
517 | misses any visible object, surface or point. An aggressive algorithm
|
---|
518 | always underestimates visibility, i.e. it never reports an invisible
|
---|
519 | object, surface or point as visible. An approximate algorithm
|
---|
520 | provides only an approximation of the result, i.e. it can overestimate
|
---|
521 | visibility for one input and underestimate visibility for another
|
---|
522 | input.
|
---|
523 |
|
---|
524 | The classification according to the accuracy is best illustrated on
|
---|
525 | computing PVS: an exact algorithm computes an exact PVS. A
|
---|
526 | conservative algorithm computes a superset of the exact PVS. An
|
---|
527 | aggressive algorithm determines a subset of the exact PVS. An
|
---|
528 | approximate algorithm computes an approximation to the exact PVS that
|
---|
529 | is neither its subset or its superset for all possible inputs.
|
---|
530 |
|
---|
531 | A more precise quality measure of algorithms computing PVSs can be
|
---|
532 | expressed by the {\em relative overestimation} and the {\em relative
|
---|
533 | underestimation} of the PVS with respect to the exact PVS. We can
|
---|
534 | define a quality measure of an algorithm $A$ on input $I$ as a tuple
|
---|
535 | $\mbi{Q}^A(I)$:
|
---|
536 |
|
---|
537 | \begin{eqnarray}
|
---|
538 | \mbi{Q}^A(I) & = & (Q^A_o(I), Q^A_u(I)), \qquad I \in {\cal D} \\
|
---|
539 | Q^A_o(I) & = & {|S^A(I) \setminus S^{\cal E}(I)| \over |S^{\cal E}(I)|} \\
|
---|
540 | Q^A_u(I) & = & {|S^{\cal E}(I) \setminus S^A(I) | \over |S^{\cal E}(I)|}
|
---|
541 | \end{eqnarray}
|
---|
542 |
|
---|
543 | where $I$ is an input from the input domain ${\cal D}$, $S^A(I)$ is
|
---|
544 | the PVS determined by the algorithm $A$ for input $I$ and $S^{\cal
|
---|
545 | E}(I)$ is the exact PVS for the given input. $Q^A_o(I)$ expresses the
|
---|
546 | {\em relative overestimation} of the PVS, $Q^A_u(I)$ is the {\em
|
---|
547 | relative underestimation}.
|
---|
548 |
|
---|
549 | The expected quality of the algorithm over all possible inputs can be
|
---|
550 | given as:
|
---|
551 |
|
---|
552 | \begin{eqnarray}
|
---|
553 | Q^A & = & E[\| \mbi{Q}^A(I) \|] \\
|
---|
554 | & = & \sum_{\forall I \in {\cal D}} f(I).\sqrt{Q^A_o(I)^2 + Q^A_o(I)^2}
|
---|
555 | \end{eqnarray}
|
---|
556 |
|
---|
557 | where f(I) is the probability density function expressing the
|
---|
558 | probability of occurrence of input $I$. The quality measure
|
---|
559 | $\mbi{Q}^A(I)$ can be used to classify a PVS algorithm into one of the
|
---|
560 | four accuracy classes according to Section~\ref{sec:accuracy}:
|
---|
561 |
|
---|
562 | \begin{enumerate}
|
---|
563 | \item exact\\
|
---|
564 | $\forall I \in {\cal D} :Q_o^A(I) = 0 \wedge Q_u^A(I) = 0$
|
---|
565 | \item conservative\\
|
---|
566 | $\forall I \in {\cal D} : Q_o^A(I) \geq 0 \wedge Q_u^A(I) = 0$
|
---|
567 | \item aggressive \\
|
---|
568 | $\forall I \in {\cal D} : Q_o^A(I) = 0 \wedge Q_u^A(I) \geq 0$
|
---|
569 | \item approximate \\
|
---|
570 | $\qquad \exists I_j, I_k \in {\cal D}: Q_o^A(I_j) > 0 \wedge Q_u^A(I_k) > 0$
|
---|
571 | \end{enumerate}
|
---|
572 |
|
---|
573 |
|
---|
574 |
|
---|
575 | \subsection{Solution space}
|
---|
576 |
|
---|
577 | \label{sec:solspace}
|
---|
578 |
|
---|
579 | The solution space is the domain in which the algorithm determines
|
---|
580 | the desired result. Note that the solution space does not need to
|
---|
581 | match the domain of the result.
|
---|
582 |
|
---|
583 | The algorithms can be classified as:
|
---|
584 |
|
---|
585 | \begin{itemize}
|
---|
586 | \item discrete,
|
---|
587 | \item continuous,
|
---|
588 | \item hybrid.
|
---|
589 | \end{itemize}
|
---|
590 |
|
---|
591 | A discrete algorithm solves the problem using a discrete solution
|
---|
592 | space; the solution is typically an approximation of the result. A
|
---|
593 | continuous algorithm works in a continuous domain and often computes an
|
---|
594 | analytic solution to the given problem. A hybrid algorithm uses both
|
---|
595 | the discrete and the continuous solution space.
|
---|
596 |
|
---|
597 | The classification according to the solution space is easily
|
---|
598 | demonstrated on visible surface algorithms: The
|
---|
599 | z-buffer~\cite{Catmull:1975:CDC} is a common example of a discrete
|
---|
600 | algorithm. The Weiler-Atherton algorithm~\cite{Weiler:1977:HSR} is an
|
---|
601 | example of a continuous one. A hybrid solution space is used by
|
---|
602 | scan-line algorithms that solve the problem in discrete steps
|
---|
603 | (scan-lines) and for each step they provide a continuous solution
|
---|
604 | (spans).
|
---|
605 |
|
---|
606 | Further classification reflects the semantics of the solution
|
---|
607 | space. According to this criteria we can classify the algorithms as:
|
---|
608 |
|
---|
609 | \begin{itemize}
|
---|
610 | \item primal space (object space),
|
---|
611 | \item line space,
|
---|
612 | \begin{itemize}
|
---|
613 | \item image space,
|
---|
614 | \item general,
|
---|
615 | \end{itemize}
|
---|
616 | \item hybrid.
|
---|
617 | \end{itemize}
|
---|
618 |
|
---|
619 | A primal space algorithm solves the problem by studying the
|
---|
620 | visibility between objects without a transformation to a different
|
---|
621 | solution space. A line space algorithm studies visibility using a
|
---|
622 | transformation of the problem to line space. Image space algorithms
|
---|
623 | can be seen as an important subclass of line space algorithms for
|
---|
624 | solving visibility from a point problems in 3D. These algorithms cover
|
---|
625 | all visible surface algorithms and many visibility culling
|
---|
626 | algorithms. They solve visibility in a given image plane that
|
---|
627 | represents the problem-relevant line set ${\cal L}_R$ --- each ray
|
---|
628 | originating at the viewpoint corresponds to a point in the image plane.
|
---|
629 |
|
---|
630 | The described classification differs from the sometimes mentioned
|
---|
631 | understanding of image space and object space algorithms that
|
---|
632 | incorrectly considers all image space algorithms discrete and all
|
---|
633 | object space algorithms continuous.
|
---|
634 |
|
---|
635 |
|
---|
636 | %*****************************************************************************
|
---|
637 |
|
---|
638 |
|
---|
639 |
|
---|
640 | \section{Summary}
|
---|
641 |
|
---|
642 | The presented taxonomy classifies visibility problems independently of
|
---|
643 | their target application. The classification should help to understand
|
---|
644 | the nature of the given problem and it should assist in finding
|
---|
645 | relationships between visibility problems and algorithms in different
|
---|
646 | application areas. The algorithms address the following classes of
|
---|
647 | visibility problems:
|
---|
648 |
|
---|
649 | \begin{itemize}
|
---|
650 | \item Visibility from a point in 3D $d({\cal L}_R)=2$.
|
---|
651 | \item Global visibility in 3D $d({\cal L}_R)=4$.
|
---|
652 | \item Visibility from a region in 3D, $d({\cal L}_R)=4$.
|
---|
653 | \end{itemize}
|
---|
654 |
|
---|
655 | This chapter discussed several important criteria for the
|
---|
656 | classification of visibility algorithms. This classification can be
|
---|
657 | seen as a finer structuring of the taxonomy of visibility problems. We
|
---|
658 | discussed important steps in the design of a visibility algorithm that
|
---|
659 | should also assist in understanding the quality of a visibility
|
---|
660 | algorithm. According to the classification the visibility algorithms
|
---|
661 | described later in the report address algorithms with the following
|
---|
662 | properties:
|
---|
663 |
|
---|
664 | \begin{itemize}
|
---|
665 | \item Domain:
|
---|
666 | \begin{itemize}
|
---|
667 | \item viewpoint (online visibility culling),
|
---|
668 | \item global visibility (global visibility sampling)
|
---|
669 | \item polygon or polyhedron (mutual visibility verification)
|
---|
670 | \end{itemize}
|
---|
671 | \item Scene restrictions (occluders):
|
---|
672 | \begin{itemize}
|
---|
673 | \item meshes consisting of convex polygons
|
---|
674 | \end{itemize}
|
---|
675 | \item Scene restrictions (group objects):
|
---|
676 | \begin{itemize}
|
---|
677 | \item bounding boxes
|
---|
678 | \end{itemize}
|
---|
679 | \item Output:
|
---|
680 | \begin{itemize}
|
---|
681 | \item PVS
|
---|
682 | \end{itemize}
|
---|
683 | \item Accuracy:
|
---|
684 | \begin{itemize}
|
---|
685 | \item conservative
|
---|
686 | \item exact
|
---|
687 | \item aggresive
|
---|
688 | \end{itemize}
|
---|
689 | \item Solution space:
|
---|
690 | \begin{itemize}
|
---|
691 | \item discrete (online visibility culling, global visibility sampling, conservative and approximate algorithm from the mutual visibility verification)
|
---|
692 | \item continuous (exact algorithm from mutual visibility verification)
|
---|
693 | \end{itemize}
|
---|
694 | \item Solution space data structures: viewport (online visibility culling), ray stack (global visibility sampling, conservative and approximate algorithm from the mutual visibility verification), BSP tree (exact algorithm from the mutual visibility verification)
|
---|
695 | \item Use of coherence of visibility:
|
---|
696 | \begin{itemize}
|
---|
697 | \item spatial coherence (all algorithms)
|
---|
698 | \item temporal coherence (online visibility culling)
|
---|
699 | \end{itemize}
|
---|
700 | \item Output sensitivity: expected in practice (all algorithms)
|
---|
701 | \item Acceleration data structure: kD-tree (all algorithms)
|
---|
702 | \item Use of graphics hardware: online visibility culling
|
---|
703 | \end{itemize}
|
---|
704 |
|
---|
705 |
|
---|