\documentclass{beamer}
\usepackage[T1]{fontenc}
\usepackage{multicol}
\usepackage{ragged2e} %new code
\usepackage[utf8]{inputenc}
\usepackage[brazil]{varioref}
\usepackage[square,sort,comma,super,authoryear]{natbib}
\usepackage{xmpmulti}
\usepackage{epsfig}
\usepackage{subcaption}
\captionsetup{compatibility=false}
\usepackage{ru,graphicx,hyperref,url} %
\addtobeamertemplate{block begin}{}{\justifying}
\setbeamertemplate{section in toc}[sections numbered]
\AtBeginSection[]
{
\begin{frame}{Table of Contents}
\begin{multicols}{2}
\tableofcontents[currentsection]
\end{multicols}
\end{frame}
}
% The title of the presentation:
% - first a short version which is visible at the bottom of each slide;
% - second the full title shown on the title slide;
\title[Modelling Cultural Heritage with Photography]{
Modelling Cultural Heritage with Photography: a New Approach for Image Acquisition using UAV}
% Optional: a subtitle to be dispalyed on the title slide
% \subtitle{Show where you're from}
% The author(s) of the presentation:
% - again first a short version to be displayed at the bottom;
% - next the full list of authors, which may include contact information;
\author[Pedro Victor Vieira de Paiva]{
Pedro Victor Vieira de Paiva\\
Marco Antônio Garcia de Carvalho\\
Eloisa Dezen-Kempter\\\medskip
{\small {pedro.paiva@pos.ft.unicamp.br}
}}
% The institute:
% - to start the name of the university as displayed on the top of each slide
% this can be adjusted such that you can also create a Dutch version
% - next the institute information as displayed on the title slide
\institute[Universidade Estadual de Campinas ]{
Laboratório de Computação Visual -- IMAGELab \\
}
% Add a date and possibly the name of the event to the slides
% - again first a short version to be shown at the bottom of each slide
% - second the full date and event name for the title slide
\date{\today}
\begin{document}
\begin{frame}
\titlepage
\end{frame}
\begin{frame}{Table of Contents}
\begin{multicols}{2}
\tableofcontents
\end{multicols}
\end{frame}
% Section titles are shown in at the top of the slides with the current section
% highlighted. Note that the number of sections determines the size of the top
% bar, and hence the university name and logo. If you do not add any sections
% they will not be visible.
\section{Motivation}
\subsection{Digital Documentation}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{Digital Documentation}
\justifying The current conditions of a construction, also called “as-is” building, must be correctly observed, obtained and analyzed in many applications \citep{remondino2011} such:
\begin{itemize}[<+-| alert@+>]
\item Historic Documentation
\item Restoration
\item Conservation
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{Digital Documentation}
\justifying Laser sensors are commonly applied in building data acquisition, resulting in an extremely precise representation, although expensive and its high computational cost.
\begin{figure}
\centering
\includegraphics[scale=.16]{images/6}
\caption{Field equipment necessary for acquisition. Source: \citep{oldow2008application}}
\end{figure}
\end{frame}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{Digital Documentation}
\begin{block}{Point Cloud}
\begin{figure}
\includegraphics[scale=.2]{images/pc}
\caption{The laser scanning process for measuring 3D point. Source: \citep{tang2010automatic}}
\end{figure}
\end{block}
\end{frame}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{Digital Documentation}
\justifying
Combining a large set of pictures, toke from different angle, and applying a method called \textit{Structure from Motion} a similar result can be achieved.
\begin{figure}
\centering
\includegraphics[scale=.7]{images/Colosseum}
\caption{Point cloud generation with SfM. Source: Bundler Project.}
\end{figure}
\end{frame}
\subsection{UAV acquisition difficulties}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{UAV acquisition difficulties}
\justifying
Unlike laser sensors, \textit{Unmanned Aerial Veichels} (UAV) with digital cameras are now affordable and user-friendly.
\begin{figure}
\centering
\includegraphics[scale=.18]{images/drone}
\caption{Eight-rotor UAV platform (BNU-D8-1) fitted with Cannon 5DII. Source: \citep{xu2014tridimensional}.}
\end{figure}
\end{frame}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{UAV acquisition difficulties}
\justifying Despite this, a universal data (image) acquisition protocol is not yet available, what make it an effective but still experimental process. Particularly in cultural heritage, is not a trivial task due:
\begin{itemize}[<+-| alert@+>]
\item Partial occlusion
\item Element uniqueness
\item Weather conditions
\item etc...
\end{itemize}
\end{frame}
\begin{frame}
\frametitle{Motivation}
\framesubtitle{UAV acquisition difficulties}
\justifying
\begin{block}{Question}
What is required to develop a standard protocol applicable to image acquisition of heritage by adopting UAV systems?
\end{block}
\end{frame}
\section{Theoretical Background}
\subsection{Photogrammetry}
\begin{frame}
\frametitle{Theoretical Background}
\framesubtitle{Photogrammetry}
\justifying
Photogrammetry is the science of making measurements from photographs, based on camera calibration.
\begin{figure}
\centering
\includegraphics[scale=.15]{images/cam_par}
\caption{Camera parameters for distance estimation. Source: \citep{wenzel2013image}}
\end{figure}
\end{frame}
\begin{frame}
\frametitle{Theoretical Background}
\framesubtitle{Photogrammetry}
\justifying
According to \citep{murtiyoso2016acquisition}, there are different established and trustworthy image acquisition protocols. This methods share common characteristics, such as:
\begin{itemize}
\item Position and sensor calibration steps
\item Angle convergence
\item Image overlay
\end{itemize}
\end{frame}
\subsection{Structure from Motion}
\begin{frame}
\frametitle{Theoretical Background}
\framesubtitle{Structure from Motion}
\justifying
\textit{Structure from Motion} (SfM) techniques uses overlapping pictures to extract object information by using camera internal parameters for orientation \citep{micheletti2015structure}.
\begin{figure}
\centering
\includegraphics[scale=.17]{images/sfm}
\caption{3D point triangulation by finding the point $\mathbf{p}$ that lies nearest to all of the optical rays $\mathbf{c}_{j} + d_{j}\mathbf{\hat{v}}_{j} $. Source: \citep{szeliski2010computer}}
\end{figure}
\end{frame}
\begin{frame}
\frametitle{Theoretical Background}
\framesubtitle{Structure from Motion}
\justifying
For outstanding outcome, it is imperative:
\begin{itemize}
\item Generous collection of images
\item Similar pictures took from rotated points of view (vertical and horizontal)
\item Depth and range variable points of view
\end{itemize}
\end{frame}
\section{Proposed Approach}
\begin{frame}{Proposed Approach}
\begin{block}{Pipeline}
\begin{figure}
\includegraphics[scale=.35]{images/dia}
\caption{Activity diagram of an effective approach for acquisition systems to structure modeling using SfM techniques. Source: author.}
\end{figure}
\end{block}
\end{frame}
\subsection{Calibration}
\begin{frame}{Proposed Approach}
\framesubtitle{Calibration}
\justifying
\begin{itemize}
\item 3D model construction requires precise sensor position estimation
\begin{itemize}
\item Start from the highest point, allowing satellite synchronization (as much as possible).
\end{itemize}
\item Regarding the camera
\begin{itemize}
\item Brightness, focus, contrast and saturation are currently well adjustable in auto-mode
\end{itemize}
\end{itemize}
\end{frame}
\subsection{Flight Protocol}
\begin{frame}
\frametitle{Proposed Approach}
\framesubtitle{Flight Protocol}
\justifying
A path capable to cover full angle variation of the structure, parallel and perpendicular, is the hardest challenge in UAV flight planning.
\begin{figure}[ht]
\begin{minipage}[b]{0.45\linewidth}
\justifying
\includegraphics[width=\textwidth]{images/fpa}
\caption{Height and weight portions fully covered.}
\label{fig:a}
\end{minipage}
\hspace{0.5cm}
\begin{minipage}[b]{0.45\linewidth}
\justifying
\includegraphics[width=\textwidth]{images/fpb}
\caption{Flight plan with high angle variation.}
\label{fig:b}
\end{minipage}
\end{figure}
\vspace{-.5cm}
\centering \small{(Source: author)}
\end{frame}
\subsection{Evaluation}
\begin{frame}{Proposed Approach}
\framesubtitle{Evaluation}
Once collected and processed, data could be evaluated comparing regional projection of point cloud to it equivalent “as-design”.
\begin{figure}
\begin{subfigure}{1\linewidth}
\begin{center}
\includegraphics[width=8cm]{images/compare_clock}
\end{center}
\end{subfigure}
\begin{subfigure}{1\linewidth}
\begin{center}
\includegraphics[width=8cm]{images/compare_wall}
\end{center}
\end{subfigure}
\caption{Comparative among projections and project views. Source: author}
\end{figure}
\end{frame}
\section{Conclusion}
\subsection{Partial Results}
\begin{frame}{Conclusion}
\framesubtitle{Partial Results}
\justifying
\begin{columns}
\begin{column}{0.47\textwidth}
\begin{figure}
\vspace{-0.3cm}
\centering
{\epsfig{file = images/elementos, width = 3.5cm}}
\end{figure}
\end{column}
\begin{column}{0.7\textwidth}
\tiny\begin{itemize}
\item[] a. Right Bell Tower;
\item[] b. Left Bell Tower;
\item[] c. Double pediment with a clock on the tympanum. Straight cymatium (2) and scrolled pediment (1);
\item[] d. Frontispiece, with spare bell towers base (cornice highlighted in the figure made up of cymatium and friezes).
\item[] g. Main door with lintel in segmental arch and door frame both in carved stone.
\item[] h. Bell towers base;
\end{itemize}
\end{column}
\end{columns}
\vspace{-0.3cm}
\tiny{
\begin{table}
\caption{Analysis of segmented regions compared to the as-designed model.}
\centering
\begin{center}
\begin{tabular}{|c|c|c|c|}
\hline
COMPONENTS & PRECISION & RECALL & ACCURACY\\
\hline
Frontispiece with voids of windows and doorways & 87,73 & 93,37 & 85,76 \\
\hline
Tower base (right) & 94,63 & 86,56 & 84,79 \\
\hline
Tower base (left) & 94,34 & 94,68 & 90,89 \\
\hline
Right Bell Tower & 85,30 & 89,45 & 83,44 \\
\hline
Left Bell Tower & 84,41 & 63,92 & 69,45 \\
\hline
Double pediment (tympanum and cymatium) & 55,68 & 39,29 & 82,32 \\
\hline
Scrolled pediment & 96,02 & 87,18 3 & 91,18 \\
\hline
\end{tabular}
\end{center}
\end{table}}
\end{frame}
\subsection{UAV and SfM popularization}
\begin{frame}
\frametitle{Conclusion}
\framesubtitle{UAV and SfM popularization}
\justifying
\begin{itemize}
\item UAV popularization and SfM algorithms allow cultural heritage documenting and modelling
\begin{itemize}[<+-| alert@+>]
\item User-friendly
\item Affordable
\item But...
\begin{itemize}
\item Still needs formalization and reliability to eliminate experimental factor
\end{itemize}
\end{itemize}
\item This work introduced a UAV image acquisition protocol able to produce strong representation for cultural heritage applications
\vspace{-.4cm}
\begin{itemize}[<+-| alert@+>]
\item Ground stations proceedings
\item SfM properties
\end{itemize}
\item Future analysis in method and application are necessary, especially with different heritage objects and SfM implementations.
\end{itemize}
\end{frame}
\appendix
\begin{frame}[allowframebreaks]{References}
\tiny
\bibliography{bib}
\bibliographystyle{plainnat}
\end{frame}
\begin{frame}[plain,c]
\begin{center}
\Huge Thank you!
\end{center}
\end{frame}
\end{document}