Non-parametric statistic modeling of SAR images based on orthogonal polynomial

For highly contaminated clutter in synthetic aperture radar (SAR) images, parametric methods are hardly effective for modeling the statistical distribution of SAR clutter. Therefore, nonparametric statistic model is suggested to solve this problem. The paper proposes a new nonparametric statistic model for SAR clutter based on the orthogonal polynomial theory. Legendre orthogonal polynomials are utilized to approximate the histogram of real SAR clutter and then CFAR detector is designed to detect ships in the SAR image. The experimental results of simulation data and real SAR data demonstrate that the proposed method is effective.