deletions | additions       diff --git a/section_Methodology_subsection_Site_Selection__.tex b/section_Methodology_subsection_Site_Selection__.tex  index 538e917..eb37e80 100644  --- a/section_Methodology_subsection_Site_Selection__.tex  +++ b/section_Methodology_subsection_Site_Selection__.tex 
...
Video data was collected at a number of sites using purpose-built video data collection systems \cite{Jackson_2013}. The video data was processed with state-of-the-art computer vision software purpose-built for traffic safety applications. The open-source software Traffic-Intelligence extracts all road user trajectories from image space using feature-based-tracking \cite{Saunier_2006}, providing cartesian coordinates of all moving objects within a scene up to thirty times a second. This high-resolution data can be analysed to provide a continuum of behavioural and traffic safety metrics across the scene including basic parameters such as speed profiles, accelerations, platoon sizes, and lane changes as well as more complex parameters such as gap (yield) time, time-to-collision, infractions, and many more surrogate safety measures. Additional software was used to partially automate the analysis (the data is quite large) and to annotate the traffic scene with contextual metadata \cite{St_Aubin_2015}.  Example data extracted from two cameras at one site is shown in Figure~\ref{fig:Trajectory_data}. The cameras were placed in the north-west corner to provide full and overlapping coverage of the roundabout. In this example, one tree did provide some occlusion of the south-east merging zone, but, as detailed in the next section, is of little importance. Such small issues of occlusion will be corrected in future versions of the software.\subsection{Analysis Zone}  At each Swedish roundabout, two cameras were positioned in such a way to cover the entire roundabout completely, and to cover at least one merging zone in each camera completely. The use of analysis zones as units of study instead of roundabouts as a whole is that, while many factors such as land use are shared, many more are not. This includes flows and flow ratios especially, but may also include geometric factors such as lane configuration, signalisation, presence of a crosswalk, approach angle, etc. which can vary from one merging zone to the next even within the same roundabout \cite{St_Aubin_2013}. Studying merging zones individually also better encapsulated the microscopic nature of the data being collected and analysed: roundabouts are often large enough for road user interactions on different sides of the roundabout to occur independently.  Table~\ref{tab:analysis_zones} lists a summary of the sites, the analysis zones at each and the geometric and land use characteristics of each.    \begin{table}  \caption{Analysis Zones Inventory}  \label{tab:analysis_zones}  \begin{tabular}{llllllll}  \hline  \textbf{Site} & \textbf{Land Use} & \textbf{Urban Density} & \textbf{Inside Radius} & \textbf{Outside Radius} & \textbf{Hourly Flow} & \textbf{Flow Ratio} & \textbf{Crosswalk} \\  \hline  SE-Fasanvagen-Trollebergsvagen-1 & Mixed & Medium & 18.0 & 25.0 & 274.8 veh/h/ln & -0.374 & Approach \& exit \\  SE-Fasanvagen-Trollebergsvagen-2 & Mixed & Medium & 18.0 & 25.0 & 241.3 veh/h/ln & -0.289 & Approach \& exit \\  SE-R103-Foretagsvagen-1 & Commercial & Very low & 15.0 & 22.0 & 237.3 veh/h/ln & 0.293 & Approach \\  SE-R103-Foretagsvagen-2 & Commercial & Very low & 15.0 & 22.0 & 182.2 veh/h/ln & 0.517 & None \\  SE-RubenRausings-Borgs-1 & Mixed & Low & 13.5 & 22.0 & 115.9 veh/h/ln & 0.646 & Approach \& exit \\  SE-RubenRausings-Borgs-2 & Mixed & Low & 13.5 & 22.0 & 82.0 veh/h/ln & 0.568 & Approach \& exit \\  SE-Svenshogs-NorraGrans-1 & Suburban residential & Low & 7.5 & 16.5 & 119.0 veh/h/ln & -0.417 & Approach \& exit \\  SE-Svenshogs-NorraGrans-2 & Suburban residential & Low & 7.5 & 16.5 & 176.9 veh/h/ln & 0.054 & Approach \& exit \\  \hline  & & & & & & &   \end{tabular}  \end{table}  The flow ratio is defined as  \begin{equation} \label{eqn:flow_ratio} Flow ratio=\frac{Q_{app}-Q_{conf}}{Q_{app}+Q_{conf}} \end{equation}  where $Q_{app}$ is the total flow rate at the approach and $Q_{conf}$ is the total conflicting flow (in the roundabout lanes at the beginning of the merging zone). It follows that a negative flow ratio would indicate that the conflicting flow is greater than the approach flow, and vice versa for a positive flow ratio.