Adaptive Detection Methods Of Moving Targets In Sea Clutter

Moving target detection in sea clutter is an important subject in the field of radar target detection. However, it is also a difficult problem because of the complex characteristic of sea clutter. Many experts and scholars at home and abroad have been dedicated to the research and exploration of sea clutter characteristics and target detection in sea clutter, and have formed a theoretical system and obtained a lot of research results. In this dissertation, the adaptive detection methods of moving targets in sea clutter are studied. A new adaptive detection method using the median power estimations and the normalized covariance estimation is proposed to improve the existing adaptive detection methods under non-uniform sea clutter environment. In order to detect low-speed moving weak targets in the case of moderate beam dwelling time, a new integration method by using the median-normalized Doppler power spectrum is developed.Firstly, there is a comprehensive review of sea clutter involved in the physical mechanism, statistical properties and statistical models. The three existing adaptive detection methods are described in detail. These methods highly rely on the covariance matrix estimation of sea clutter. Several commonly-used covariance matrix estimation methods are reviewed. Aiming at inhomogeneous sea clutter environment, a new estimation method, which depends upon the median power estimate and normalized covariance matrix estimate in reference range cells, is proposed. In this method, the clutter power at the reference cells firstly are normalized to one unit and then the normalized covariance matrix is estimated from the normalized secondary data at the reference cells. The median power at the reference cells is used as the power estimate of the sea clutter at the cell under test(CUT). As a result, the product of the median power estimate and normalized covariance matrix estimate forms an efficient estimate of the covariance matrix of sea clutter at the CUT. The new estimator can effectively mitigate the impact of the power fluctuation of sea clutter along range cells. Replacing the clutter covariance matrix estimates by the new estimate in adaptive detectors yields new adaptive detectors that can improve the performance in inhomogeneous sea clutter environments. The performance improvement is verified by real sea clutter data and the results show that using the new clutter covariance matrix estimate brings obvious merit in inhomogeneous sea clutter environment relative to using the traditional sample covariance matrix(SCM) estimates.Secondly, the concept and methods of the power spectrum estimation of sea clutter time series is introduced. The characteristics of sea clutter Doppler power spectrum is analyzed and the analysis shows that the power of sea clutter at individual Doppler bins are non-uniform, meaning different power means and variances. In addition, in the case of moderate beam dwelling time from tens of millisecond to several seconds, the Doppler offset of a moving target often distribute into several adjacent Doppler bins. The ratio of the Doppler power spectrum at the CUT to the median Doppler power spectrum onto the reference range cells forms the median-normalized Doppler power spectrum of the CUT. An new integrator of the median-normalized Doppler power spectrum along the Doppler bins is developed for low-speed moving weak target detection. The proposed detector is verified by real sea clutter data and the results show that it obtains better detection performance than the recent fractal-based detection method.