Detection background segmentation and statistical analysis for High Frequency Surface Wave Radar

A scheme of detection background segmentation and statistical analysis for High Frequency Surface Wave Radar (HFSWR) is presented. Clutter and noise in HFSWR is multi-sourced of which statistical distribution is complex. To overcome this problem, a detection background segmentation algorithm based on the cloud model is proposed to divide the background into several clutter/noise regions, and Weibull distribution is used to fit the segmentation regions, noted as seg-regions. According to the statistical fitting results, the difference of seg-regions is measured by KL divergence, and the segmentation algorithm is optimized.