CLUTTER REDUCTION IN SYNTHETIC APERTURE RADAR IMAGES WITH STATISTICAL MODELING: AN APPLICATION TO MSTAR DATA

摘要

In this article, an application of clutter modeling and reduction techniques to synthetic aperture radar (SAR) images of moving and stationary target acquisition and recognition data is presented. Statistical modeling of the clutter signal within these particular SAR images is demonstrated. Lognormal, Weibull, and K-distribution models are analyzed for the amplitude distribution of high-resolution land clutter data. Higher-order statistics (moments and cumulants) are utilized to estimate the appropriate statistical distribution models for the clutter. Also, Kolmogorov-Smirnov (K-S) goodness-of-fit test is employed to validate the accuracy of the selected models. With the use of the determined clutter model, constant false-alarm rate detection algorithm is applied to the SAR images of several military targets. Resultant SAR images obtained by using the proposed method show that target signatures are reliably differentiated from the clutter background.