Feature extraction in speckled imagery using dynamic B-spline deformable contours under the G(0) model

Abstract

This article proposes a technique for boundary detection in speckled imagery based on active contours and on the statistical properties of speckled data. Synthetic Aperture Radar (SAR), sonar, B‐ultrasound and laser imagery is corrupted by a signal‐dependent non‐additive noise called speckle. Many statistical models have been proposed to describe this noise, aiming at the development of specialized techniques for image improvement and analysis. The distribution is a statistical model that succeeds in describing a wide range of areas as, for instance in SAR data, pastures (smooth), forests (rough) and urban (extremely rough) areas. In order to determine boundaries between different areas, dynamic B‐splines are proposed in this paper. The algorithm automatically finds seeds and regions boundaries by a decision rule based on the estimation of the parameters of the law. The proposed algorithm is tested on synthetic and real SAR images, and both the accuracy and the performance of the proposal are assessed.