Zuzana KRIVÁ

Basic principles in SAR imagery filtration in remote sensing

Krivá, Z.: Basic principles in SAR imagery filtration in remote sensing. Kartografické listy, 2016, 24 (2).

Abstract: SAR (Synthetic Aperture Radar) is a radar with its antenna prolonged in a “synthetic way” – by a software to achieve better resolution. SAR imagery used in remote sensing suffers from a heavy noise called speckle which degrades the quality of images and makes their interpretation difficult, thus filtering must be the first step in their interpretation and to develop applications. The received signal is given by the coherent (vectorial) sum of the signals received from each individual scatterer in the resolution cell – we cannot trust in a single value. The solution is the averaging. The filters are required to average homogenous areas and to preserve details. The speckle is a multiplicative noise, i.e. the larger is the reflectivity, the larger is the noise. In this paper we present the basic principles of statistical filter and we show their relationship to filters based on diffusion partial differential equations (PDEs).

Keywords: synthetic aperture radar, speckles, additive noise, multiplicative noise, statistical filter, diffusional partial differential equations