AUTOMATED ANALYSIS AND FUSION OF MULTI-SPECTRAL DATA IN AIRBORNE OIL SPILL REMOTE SENSING*

摘要

During the last thirty years, aerial surveillance of marine oil spills has become a multinational effort, especially for the deterrence of potential polluters and support of oil spill clean-up crews. For many years, the main effort has been directed towards developing sensors with enhanced spill monitoring capabilities which explains the existing large number of wellestablished oil spill remote sensors. Recently, more attention has been paid to the automated derivation of high-level information from airborne remotely sensed multi-spectral oil spill data. In this paper, we present an algorithmic framework for automated analysis and fusion of multispectral oil spill data acquired by the German maritime surveillance aircraft. It is shown how infrared/ultraviolet, microwave radiometer and imaging laser fluorosensor data can be converted into high-level information using single-sensor data processing and multi-sensor data fusion. It is also explained how this can improve the usability of the maritime surveillance system. In this context two concepts are introduced: (1) the internal data processing and (2) the external information processing. While the internal processing of pollution data is spatially limited to the surveillance aircraft, the external information processing includes the transfer of the internally derived high-level information products to ground stations and aspects such as GIS integration and AIS communication. This paper describes the automated internal data processing in terms of analysis and fusion of multi-spectral oil spill data and outlines opportunities of external information processing.