Assessment of speckle reduction filters for automatic detection of small/thin features [in SAR images]

Abstract

Evaluates the effectiveness of several speckle reduction filters in extracting small/thin features in SAR images. The filters were tested on various types of bright and dark features in observed images instead of simulated images. The filters were required to preserve contrast between the feature and its immediate background, as well as reduce speckle in the image. Based on the results of the tests, the authors found that the Small Feature Preserving (SFP) filter was best for extracting bright features, and that the Sigma filter was best for extracting dark features.