A filter for reducing speckle in synthetic aperture radar images
Nobuyoshi Kobayashi, H. Hirosawa · Electronics and Communications in Japan (Part I Communications) · 1996
Abstract A filter for reducing the speckle in synthetic aperture radar (SAR) images is proposed. This filter is based on the most likelihood method and consists of the following procedures. First, by using ann X nwindow, a spatial averaging is carried out and two‐dimensional data composed of sets of average value μ and variance o2are generated. Next, by also applying then X nwindow, the samples, without being affected by edges are extracted among the samples contained in the window based on a criterion derived from the statistical characteristic of the speckles. For the extracted samples, the average value of μ is computed and the derived value is given as the output for the pixel at the center of the window. For this filter, smoothing characteristics and the operation at an edge are evaluated by simulation. Applying the filter to the SAR images allows it to be confirmed that the filter operates correctly.