Kullback-Leibler divergence based range-Doppler spectrum region partition method for OTHR

A Doppler-range spectrum region segmentation method for Over-the-Horizon Radar (OTHR) is proposed as a preprocessor for the following signal processing. The method derives from the well-known Kullback-Leibler (KL) Divergence which measures dissimilarity between the sea clutter and atmosphere noise. By thresholding the bias of KL map it is estimated by a statistical and variational multiphase level set method. Compared with the Otsu method and the deviation based method by experiments used synthetic data and real data, the proposed algorithm shows a better performance which is robust to the interference of targets and speckled noise.