Multiresolution Hierarchical Target Detection in High-resolution Maritime Surveillance Radars

High-resolution maritime surveillance radars need to detect small targets and moderate/large targets such as ships in short/medium distance regions. At a single range resolution, radar echoes of moderate/large targets severely affect detection of small targets around them. In this paper, a simple digital filtering method is given to realize range resolution conversion by which multiresolution radar echoes data are generated from high range resolution data. As one of the main contributions, the compound-Gaussian model with inverse Gaussian textures (CGIG) is extended to characterize multiresolution sea clutter. Moreover, an across -resolution parameter estimation method is proposed to estimate the parameters of the multiresolution CGIG model based on across-resolution moment relationship. The other contribution is a hierarchical target detection scheme in the multiresolution CGIG clutter model. Simple non-adaptive noncoherent integration detectors are used at lower resolution channels to detect moderate/large targets and adaptive near-optimum coherent detectors in the CGIG distributed sea clutter are adopted at higher resolution channels to detect small targets using the detections from the lower resolution channels as the prior information on environment. Finally, the multiresolution CGIG model and hierarchical target detection scheme are examined by an X-band island-based measured data using an unmanned aerial vehicle (UAV) as test target.