Search

As a basic characteristic of sea clutter, the reflectivity of sea surface depends on many factors. Various universal empirical models of low precision have been developed to predict the reflectivity of sea surface. In this letter, a method is proposed to train specific predictors by big data learning, where the universal empirical models are embedded to the architecture of the generalized regression neural network (GRNN) to enhance the learning ability and efficiency. On the sea clutter database measured by an island-based UHF-band radar in the offshore waters of the Yellow sea of China at low grazing angle, the GRNN-based predictors of different structures are compared with other predictors. The results on the database show that the GRNN-based predictors behave better at learning efficiency, prediction precision, and robustness.

Reflectivity is an important characteristic of sea clutter that depends on the parameters of the radar and the environment. This paper presents reflectivity measurements of UHF-band horizontal pola...

Wind direction factor, as a part of some empirical sea clutter models of the normalised radar cross-section (NRCS), is closely related to wind direction angle, grazing angle and radar wavelength. The use of the low-grazing-angle (LGA) sea clutter measured by UHF-band radar on the Lingshan island to discuss the characteristic of the wind direction factor is investigated. The results indicate that the upwind-to-crosswind ratio of NRCS, which is calculated by the empirical model, could not provide a good fit for grazing angles

In this paper, One-Class SVM (OCSVM) detector based on three detection features is proposed to distinguish floating small targets from sea clutter. Normalized Hurst exponent, relative Doppler peak height and relative Doppler vector entropy are extracted from radar returns as three detection features. Then, the separating hyperplane of OCSVM detector attending false alarm rate requirement is obtained by kernel trick of SVM. Therefore, the detection decision could be made according to the position of returns' features, which is familiar with outlier detection. The results of simulation experiments based on measured data sets show that the proposed detector acquires better detection performance compared with some existed detectors.