Flycapture: Used with the BB2 camera to communicate with and captura data.
                    see http://www.ptgrey.com/flycapture-sdk

        Triclops: The other side of flycapture, that turns the images taken from
                  the camera and turns them into 3d points.
                  see http://www.ptgrey.com/triclops

        Point Cloud Library: Used to hold and manipulate the point cloud objects.
                                Extremely useful and well documented, all the data manipulation
                                and processing goes on in here.
                                see http://pointclouds.org/documentation/tutorials/

                              An example of all three of these technologies to take and then display
                                a point cloud image from the camera can be found here: https://github.com/sterlingbaldwin/IGV_code/blob/master/bb2cloud/bb2viewer.cpp

                              Get image data -> Create empty visual field grid model -> Analyse the image to get obstacle locations
                                and place non-passable entries in the model. Once all non-passable locations have been identified,
                                find the current target location in the grid -> Generate path to target -> Execute path

                              My main plan here is to grab the cloud image from the Triclops and put it directly into
                                a pcl cloud. From there we need to find two things: 1) the white lines denoting walls,
                                and 2) barrels/flags/cone obstacles to avoid and/or direct our movement within the allowed
                                space. To find the lines, first do a planar segmentation pass to remove all non-ground
                                points from the image, then do a pass of luminosity filtering removing all points but the
                                top 10% of whiteness points. Then we use the lines to mark their corresponding location
                                in the field model as non-passable.

                              From the planar segmentation, we take the outliers, ie. the points that dont belong to the
                                  ground (and then must be cones/barrels/flags) check if their flags, and otherwise also
                                  place their corresponding points to the non-passable set.

                              Once we have the location of all the obstacles accounted for, we mark the grid location
                                  that is both the closest to the destination, but also adjacent to unexplored locations.
                                  When we have our short term target, then we can create a minimum-spanning tree and
                                  follow its branch to the target node.


        Galil: The controller is set to a static IP address of 192.168.1.3. The controller can be
                communicated directly over its socket interface, which we do with a simple python
                script. See https://github.com/sterlingbaldwin/IGV_code/blob/master/utils/girshell.py