DIGITAL

Scientific publication

Applying image processing methods for correction of weather radar data

Publication from Digital

Kaltenböck, Rudolf and Croonen, Gerardus and Ganster, Dr Harald Ganster, Hennermann, Karin and Kerschbaum, Markus and Nowak, Christoph and Mayer, Stefan and Steginska, Sylwia and Uray, Martina

Proc. 6th European Conf. on Radar in Meteorology and Hydrology: Adv. in Radar Technology (ERAD 2010) , 1/2010

Abstract:

High resolution spatial and temporal radar data are used extensively for nowcasting and warnings in aeronautical meteorology and air traffic management. For dealing with this in time procedures, radar data will be interpreted by forecasters, automatically processed for further applications and as well as delivered to end-users for high end user products. The later quality and reliability highly depends on pre-processing and correction of non-meteorological echoes. There is a great variety of different sources for artefacts and errors in weather radar data like targets from ground clutter or biological/aircrafts, interferences to external emitters (e.g. RLAN, other radars, sun) or blocking and attenuation. During the past decade in Europe an increase of spurious
 signals on weather radar data has been detected and subsequent adequate filters to be in demand. After multi-temporal/multi-parameter ground-clutter elimination, the maximum projected reflectivity product is corrected by applying digital image processing methods. The digital image algorithms are able to cope with the high variety of artefact appearance and allow for reliable detection of artefacts and uncovered zones within radar images. The following results will be presented: a) The detection
 of artefacts with the aid of texture analysis and pattern recognition and trend analysis of images show reliable results for removal of radial and spherical artefacts and show b) potential for correction of attenuation without use of dual polarisation data. c) For correction of blocking areas, additional satellite data will be combined by applying pattern recognition and classification approaches to fill the gap areas of radar data.