Feature detection in synthetic aperture radar images using fractal error

Abstract

This paper discusses a technique that will enhance man-made features in SAR images. The technique uses a metric called fractal error. Developed by Cooper, et al. (1994) for aiding photointerpreters in detecting man-made features in aerial reconnaissance images, this metric is based upon the observed propensity of natural image features to fit a fractional Brownian motion (fBm) model. Natural scene features fit this model well, producing a small fractal error. Man-made features, on the other hand, usually do not fit the fBm model well and produce a relatively large fractal error. Therefore the fractal error is useful as a discriminant function for detecting man-made features in SAR imagery. The fractal error metric is defined, an approach to segmentating man-made objects in SAR images is discussed, and the results are presented.