As mentioned in paragraph 3.2, when obstacle detection means the mere localization of objects that can obstruct the vehicle's path without their complete identification or recognition, stereo IPM can be used in conjunction with a geometrical model of the road in front of the vehicle. Assuming the flat road hypothesis introduced in the previous section, IPM is performed using the same relations. This is of basic importance since in a system aimed to both obstacle and lane detection the IPM transform can be performed only once and its result can be shared by the two processes. The flat road model is checked through a pixel-wise difference between the two remapped images: in correspondence to a generic obstacle in front of the vehicle, namely anything rising up from the road surface, the difference image features sufficiently large clusters of non-zero pixels that have a particular shape.
Due to the different angles of view of the stereo cameras, an ideal homogeneous square obstacle produces two clusters of pixels with a triangular shape in the difference image, in correspondence to its vertical edges.
Unfortunately triangles found in real cases (see figure 4) are not so clearly defined and often not clearly disjoint because of the texture, irregular shape, and non-homogeneous color of real obstacles. Nevertheless clusters of pixels having an almost triangular shape are anyway recognizable in the difference image (see figure 4.e). The obstacle detection process is thus based on the localization of these triangles.
Moreover, this process is complicated by the possible presence of two or more obstacles in front of the vehicle at the same time, thus producing more than one pair of triangles, or partially visible obstacles, thus producing a single triangle; a further processing step is thus needed in order to classify triangles that belong to the same obstacle.