A novel processing chain ingesting multi-band SAR data for fully automatic oil spill detection

With the launch of the Italian constellation of small satellites for the Mediterranean basin observation (COSMO)-SkyMed and the German TerraSAR-X missions, the availability of very high-resolution SAR data to observe the Earth day or night has remarkably increased. Moreover, future SAR missions such as Sentinel-1 are scheduled for launch in the very next years. Suitable and adequate image processing procedures are then necessary to fully exploit the huge amount of data available. As far as oil spill detection is concerned, it is known that one of the most critical issues for the implementation of a fully automatic processing chain is the image segmentation. In fact, the extraction of the dark spots in the image is the first of three necessary steps, the other two being its characterisation by using a set of features and the classification between oil spill and look-alike. Different approaches have been discussed in the literature for dark spot detection in SAR images acquired over the oceans. Aside from the accuracy of the segmentation results, one of the most significant parameters for evaluating the performance in this context is the processing time which is necessary to provide the segmented image. This might be crucial both in emergency cases and when stack of images have to be elaborated in a raw. In this paper we present a new fast, robust and effective automated approach for oil-spill monitoring. A combination of Weibull Multiplicative Model (WMM), Pulse Coupled Neural Network (PCNN) and Multi-Layer Perceptron (MLP) techniques is proposed for achieving the aforementioned goals. One of the most innovative ideas is to separate the detection process into two main steps, WMM enhancement and PCNN segmentation