A New Method for Ship Detection in SAR Imagery Based on Combinatorial PNN Model

The probabilistic neural network (PNN) model plays a very important role for ship detection in synthetic aperture radar (SAR) imagery, however there are still some detection parameter need to improve for the requirement of detection accuracy and speed. This paper presents a new method based on combinatorial PNN model for ship detection in SAR imagery. The method includes 8-bit and 16-bit image processing models, and an improved probabilistic neural network model is proposed, a new constant false alarm rate (CFAR) calculation algorithms is adopted. Compared with convention PNN-based ship detection method, the new method based on combinatorial PNN model performs well.