Multi-channel Maritime Radar Target Detection using Morphological Component Analysis

Radar detection of small maritime targets can be difficult due to the dynamic motion of the sea. In this paper, a new space-time signal separation technique is presented which can extract both stationary and moving targets from the background interference. This is achieved with a space-time implementation of the morphological component analysis using a two dimensional short time Fourier transform as the dictionary. An analysis of the choice of penalty parameter is also presented to ensure good separation of the target and interference components. The algorithm performance is demonstrated using a Monte-Carlo simulation with synthetic multi-channel sea clutter data modelled on the Ingara X-band sea clutter data set and the processed output tested against a number of practical detection schemes.