deletions | additions       diff --git a/section_Methodology_subsection_Site_Selection__.tex b/section_Methodology_subsection_Site_Selection__.tex  index 9b088fc..a8963f3 100644  --- a/section_Methodology_subsection_Site_Selection__.tex  +++ b/section_Methodology_subsection_Site_Selection__.tex 
...
One important distinction between the Québec and Swedish roundabouts are the pedestrian and, especially, cyclist flows. Cyclist flows in the Québec roundabouts are virtually non-existent; however these flows are non-trivial at many, but not all Swedish roundabouts. Consequently, only Swedish roundabouts with limited pedestrian or cyclist flows were considered for this study (low pedestrian and cyclist flows are still very common in Swedish roundabouts in low-density areas). It's also worth noting that besides aesthetic differences, the design of the selected Swedish roundabouts has one vital difference with Québec roundabouts: virtually all sidewalks in the town of Lund integrate a cycle lane, including at crosswalks. On the other hand, it is common for these sidewalks to bypass the roundabouts entirely via tunnels \citep{Sakshaug2010}. Given this, cycling \textit{within} the roundabout is rare both in Québec and at the Swedish sites selected.  \subsection{Behavioural Data} \subsection{Traffic Data Collection}  Video data was is  collected at a number of sites using purpose-built video data collection systems \cite{Jackson_2013}. The video data was is then  processed with state-of-the-art computer vision software purpose-built for surrogate safety analysis 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 thirty times a second. This high-resolution datais intended for analysis to provide a set  ofbehavioural data and surrogate safety measures between  all road users within a traffic scene. The parameters of interest scene is needed  forthis particular study are speed (and speed profiles), time-to-collision (TTC) \citep{Hayward_1971}, and yielding post-encroachment time (yPET), an ordinary PET measure \citep{allen1978analysis} measured specifically at  the yield line. Additional road user behaviour analysis to follow. Given the large volume of data to be processed, additional  software was is  used to partially calibrate and  automate the analysis and to annotate the traffic scene with contextual metadata \cite{St_Aubin_2015}. Example trajectory  data extracted from two cameras at one site is shown in Figure~\ref{fig:Trajectory_data}. In this case two cameras are placed in the north-west corner to provide full and overlapping coverage of the roundabout.