Multipixel Optical Detection of the Weak Floating Objects on an Agitated Sea Surface

Abstract

In this work, a new detection algorithm is proposed to improve the well-known Matched Subspace Detector (MSD) in a new scenario where the object is weak and with a priori unknown parameters. Detection algorithms using statistical parameters MSD are designed in base of the General Likelihood Ratio Test (GLRT), because they can provide a universal solution to detection problems where object parameters are partially or totally unknown. However, many works and papers to today were proposed minimum changes to this detector in order to adapt it to specific environments and improve the performance of the same detector. Both algorithms are implemented in simulated environments and real video cases in order to verify the improvement of the new algorithm.