Laser-Based Obstacle Detection at Railway Level Crossings

Figure 7

An illustration of the process employed to determine which detected objects represent the half-barriers (HB) in the point cloud. The coordinate system represents the 2D projection of the LIDAR frame of reference. The filled circles correspond to detected objects described in terms of the LIDAR frame of reference, that is, putative HB. (a) and (c) Four putative HB were randomly selected to form quadrilateral , whose centroid is . (b) and (d) The quadrilaterals picked in (a) and (c) are centred in the origin by translating each vertex by the corresponding quadrilateral’s centroid vector. After this transformation it is possible to compare each putative HB position, a given vertex , with the corresponding expected position (unfilled circles). After averaging these distances, it is possible to state that the hypothesis picked in (a) is most likely to be correct one when compared to the one picked in (c).