Target Recognition by Means of 3D ISAR Images

In this paper we will compare two different target recognition algorithms based on the exploitation of 3D Interferometric ISAR (InISAR) in a maritime environment. The first algorithm uses a model-based approach benefiting from a 3D CAD database, while the other uses a silhouette-based approach based on 2D shapes of the target model. This goes beyond the simple classification of the general type of target, going deeper into the recognition of targets among the same type of class (a database of ships). Real data acquired during two measurement campaigns with a swarm of drones and a system emulating a shipborne platform will be used. The recognition outcome of the two methodologies will be compared to provide their strength and limitations.