Permalink
Browse files

changed slides based on Hussein's recommendations

  • Loading branch information...
timminata committed May 23, 2012
1 parent a3535e3 commit 8d4ddc7eaa03412bfadd2f4917ab1d510783370d
Binary file not shown.
Binary file not shown.
@@ -70,13 +70,14 @@ \subsection{}
\begin{columns}
\begin{column}{.65\linewidth}
\begin{itemize}
- \item Started by the UCT Department of Geomatics in 2004
- \item Aims to preserve cultural heritage
+ \item Started by the UCT Department of Geomatics
+ \item Aims to preserve cultural heritage by documenting heritage sites
+ \item Documenting a site involves producing:
\begin{itemize}
- \item Plans
- \item Laser scans
- \item Photos
- \item Videos
+ \item Detailed Laser scans
+ \item Photographs
+ \item Video Tours
+ \item Geographical Information Systems
\end{itemize}
\end{itemize}
\end{column}
@@ -92,9 +93,10 @@ \subsection{}
\begin{columns}
\begin{column}{.65\linewidth}
\begin{itemize}
- \item Over 100 models of sites in various African countries
- \item These models are big
- \item Containing upwards of a billion points
+ \item To create the model a point cloud is generated using a laser scanner
+ \item The project has produced close to 100 models of sites in various African countries
+ \item Some of the models are very large
+ \item Many contain over a billion points
\end{itemize}
\end{column}
\begin{column}{.35\linewidth}
@@ -105,11 +107,11 @@ \subsection{}
\end{frame}
-\begin{frame}{Problems Faced}
+\begin{frame}{Problems Faced by the Geomatics Department}
\begin{itemize}
\item Fast-growing volume of data
\item Difficulty in storing the data
- \item Difficulty with viewing the large models in real-time
+ \item Difficulty with viewing the large models
\item Data management issues
\item Data locality issues
\end{itemize}
@@ -118,7 +120,7 @@ \subsection{}
\begin{frame}{Problems Approached}
\begin{itemize}
- \item Difficulty with viewing the large models in real-time
+ \item Difficulty with viewing the large models
\item Data management issues
\end{itemize}
\end{frame}
@@ -147,7 +149,7 @@ \subsection{}
\section{Dynamic Point Cloud Streaming}
\subsection{Tim Trewartha}
-\begin{frame}{Dynamic Viewing of Large 3D Models}
+\begin{frame}{Dynamic Point Cloud Streaming}
\begin{itemize}
\item The Department of Geomatics has indicated that they have difficulties
handling the sizes of some of their models
@@ -158,36 +160,31 @@ \subsection{Tim Trewartha}
\end{itemize}
\end{frame}
+\begin{frame}{Current System}
+\includegraphics[scale=1.4]{images/laser.jpg}
+\end{frame}
-\begin{frame}{Research Question}
-\begin{itemize}
-\item Is it feasible to support real time viewing of models containing
- billions of points?
-\item Answering this question in the affirmative would enable exploration
- of the Zamani models in their full detail
-\item It would have a significant impact in the Geomatics department
-\end{itemize}
+\begin{frame}{Proposed System}
+\includegraphics[scale=1.4]{images/laser2.jpg}
\end{frame}
-\begin{frame}{Proposed Solution}
+\begin{frame}{Research Questions}
\begin{itemize}
-\item Implement a multi-resolution data structure to divide our model into
- manageable chunks
-\item Initially only a subset of the points is available
-\item As one zooms into the model and greater detail is required, we dynamically
- fetch additional points from our data structure until the full original
- detail is available
+\item Is it feasible to support viewing of models containing
+ billions of points at interactive frame rates?
+\item Can the level-of-detail data structure implemented facilitate streaming of
+ the models from a central server to clients?
+\item Answering these questions in the affirmative would enable exploration
+ of the Zamani models in their full detail
+\item It would have a significant impact in the Geomatics department
\end{itemize}
\end{frame}
-
\begin{frame}{Relevant Literature}
\begin{itemize}
-\item Common approaches to structuring large 3D models include octrees,
- R-trees, bounding sphere hierachies, and Hilbert Space Filling
- Curves
+\item Common approaches to structuring large 3D models include octrees and bounding sphere hierachies, as well as many others.
\end{itemize}
{\scriptsize
\begin{tabular}{l|c|l}
@@ -203,27 +200,43 @@ \subsection{Tim Trewartha}
\end{frame}
+\begin{frame}{Proposed Solution}
+
+\begin{itemize}
+\item Implement a multi-resolution data structure to divide our model into
+ manageable chunks
+\item Initially only a subset of the points is available
+\item As one zooms into the model and greater detail is required, we dynamically
+ fetch additional points from our data structure until the full original
+ detail is available
+\item The actual data can either be stored locally or streamed from a server
+\end{itemize}
+\end{frame}
-\begin{frame}{Proposed DataStructure}
+
+\begin{frame}{Data Structure}
\begin{itemize}
-\item From researching the literature it seems that octrees have the best
+\item From the literature it seems that octrees have the best
performance
\item All data is stored in the leaf nodes
\item Inner nodes provide simplified multi-resolution representations
\item No leaf node should contain more than a specified number of points
-\item Empirically, it seems that a value of around 30,000 gives good performance
+\item Empirically, it seems that a value of around 30,000 points gives good performance
\end{itemize}
\end{frame}
+\begin{frame}{Octree DataStructure}
+\includegraphics[scale=0.35]{oct.png}
+\end{frame}
+
\begin{frame}{Evaluation Criteria}
\begin{itemize}
-\item Can the system render the largest of the Zamani models at interactive
- frame rates?
+\item Has the hierarchical data structure successfully facilitated level-of-detail
+ streaming of the Zamani models?
+\item We will evaluate this by testing multiple models of varying sizes
\item If this goal is achieved the system will be a success
-\item Varying degrees of success can also be determined by testing smaller
- models of varying sizes
\end{itemize}
\end{frame}
}
@@ -266,7 +279,7 @@ \subsection{Michiel Baird}
\begin{itemize}
\item How effective is an automated workflow solution in Geomatics context?
\item Can data management be implemented using a workflow system?
-\item This would aid in reproducability of the research as well
+\item This would aid in reproducibility of the research as well
as address the data management issues
\end{itemize}
\end{frame}

0 comments on commit 8d4ddc7

Please sign in to comment.