Removal of spurious radar echoes with a Meteosat neural network precipitation classifier
Pankiewicz, G. ; Johnson, C. ; Harrison, D.
This report describes the development of a new neural network scheme based on the use of Meteosat infrared, or visible and infrared imagery, to determine a probability of precipitation within the Nimrod domain, for use in removing spurious echoes from Nimrod radar composite images. The neural networks were trained on samples taken from 3200 boxes of 17x17 pixels, selected from 48 sets of Meteosat and radar images, taken throughout the period July 1995 to June 1997 during day or night. Feature selection was performed on the images to look at the ability of various visible and infrared features calculated over different sized regions to discriminate rain from no rain at a threshold of 1/32 mmh to power -1. The Meteosat features which were found to best discriminate rain from no rain in individual 5km radar pixels were the central value, the minimum, maximum range and the ratio of maximum to minimum of 7x7 pixel infrared samples. When visible reflectivity data were also available, the best features were found to include the same features for the visible samples, except the central value. Two ...
Accès à la notice sur le site du portail documentaire de Météo-France