A robust constant false alarm rate detector based on the Bayesian estimator for the non-homogeneous Weibull Clutter in HFSWR

Target detection is a challenging problem in the complex detection environment of high-frequency surface wave radar (HFSWR). Due to the existence of various clutter and interferences, the detection environment in HFSWR is highly non-homogeneous with sharp/smooth clutter edges and multiple interfering targets. In this paper, we propose an effective detector for the non-homogeneous Weibull clutter in HFSWR with the use of prior information on the environmental context. Considering the spatial information of distribution parameters among neighboring cells and sparse information of interfering targets, the proposed method uses the Bayesian method to estimate distribution parameters and designs an objective function, which is composed of the data fidelity term, block matching 3-D frames and a sparsity regularization term. With the estimates of distribution parameters, the detection threshold is calculated accordingly and the presence of the target cell is decided. By minimizing the proposed model, the proposed method provides robust and accurate estimates of distribution parameters and can achieve the root-mean-squared error less than 0.12 in the simulated detection scenarios. Simulation and experimental results from a real HFSWR system are given to demonstrate the effectiveness of the proposed detector.