RFI detection algorithm: Accurate geolocation of the interfering sources in SMOS images

This paper presents the algorithm developed by the SMOS team to accurately detect the coordinates of the Radio Frequency Interferences (RFIs) sources observed by SMOS. This information is provided then to the national frequency management authorities, as part of a global strategy to improve the RFI situation in the SMOS data. The algorithm makes use of the large amount of observations from SMOS in order to increase the geo-location accuracy of the on-ground emitters despite SMOS moderate spatial resolution for this specific application. On top of that, some emissions detected from several interferences world-wide are so strong that the secondary side-lobe levels associated to the point-source type of signal are much larger than the expected natural radiation from Earth. In this conditions, the detection of the real location of the emitter becomes an important challenge. This problem is overcome with a set of data filters related to the specificities of SMOS reconstructed images. More than 150 RFIs have been switched off thanks to the information provided by this detection algorithm.