A Rotation-Invariant Transform for Target Detection in SAR Images

Rotation of targets poses a great challenge for the design of an automatic image-based target detection system. In this paper, we propose a target detection algorithm that is robust to rotation of targets. Our key idea is to use rotation invariant features as the input for the classifier. For an image in Radon transform space, namely R(b, θ), taking the magnitude of 1-D Fourier transform on θ, we get |Fθ{R(b, θ)}|. The rotation invariance of the coefficients of the combined Radon and 1-D Fourier transform, |Fθ{R(b, θ)}| was proved in this paper. These coefficients are used as the input to a maximum-margin classifier based on I-RELIEF feature weighting technique. The objective of the I-RELIEF technique is to maximize the margin between two classes and improve the robustness of the classifier against uncertainties. For each pixel of the Synthetic Aperture Radar (SAR) image, a feature vector can be extracted from a sub image centered at that pixel. Then our maximum-margin classifier decides whether the pixel is target or non-target which produces a binary-valued image. We further improved the detection performance by connectivity analysis, image differencing, and diversity combining. Our performance evaluation of the proposed algorithm was based on the data set collected by Swedish CARABAS-II systems. In conclusion, the experimental results show that our proposed algorithm achieved superior performance over the benchmark algorithm.