The search for methods of small targets detection by marine radars is carried out in this paper. As a result of the sources analysis three promising group of methods are selected – adaptive methods based on the clutter amplitude spherically invariant random process representation, methods based on non stohastic clutter representation and methods based on orthogonal transformations. Separately, neural networks and clutter polarization properties are noticed. Most of researches consider compound Gaussian distribution for high resolution sea clutter data modeling. Adaptive CFAR algoritm developed by Kelly for Gaussian clutter was extended for spherical invariant random process. This algorithm can detect target in compound Gaussian clutter and ensure CFAR detection if covariance matrix is precisely known. Many researches proposed different methods of covariance matrix estimating using information from cells adjacent with cell under test. Progress in this field is moving to finding less computation cost but more precise methods of covariance matrix estimation. Further progress for small slow moving targets in heavy sea clutter detection can develop in 3 ways – improving covariance matrix estimation, exploiting not stochastic methods and finding best orthogonal transform for sea clutter Doppler spectrum describing.

В работе осу-ществляется поиск методов обнаружения малоразмерных малоподвижных целей мор-скими РЛС. В результате анализа источников избирается три перспективные группы методов – адаптивные методы, основанные на представлении амплитуды клатера сферически инвариантным случайным процессом, методы, основанные на моделирова-ния клатера детерминированным процессом и методы, основанные на ортогональных преобразованиях. Отдельно отмечается использование нейронных сетей и поляриза-ционных свойств клатера.