While the modelling and features extraction step of the algorithm, the system search for identification criteria on all the moving objects observed. There is a tracking task at the begining of this step, consisting of determining if the observed object were present in the previous frames. If not, a new identification subject is created and the triangulation of its 3D shape is done. If the object were present in the previous frames, the system also triangulate its shape but it also have to registrate it with earlier observation that were done before to obtain a better model. At the end of this step, the system store the maximum motion speed, the size and the 3D shape of the object.