Ocean Surface Wind Direction Inversion from HFSWR Based on BP Neural Network

Ocean surface wind direction information can be extracted from high frequency surface wave radar (HFSWR) sea echo's positive and negative first-order spectra. In this paper, BP neural network (BPNN) model is proposed to inverse the wind direction from HFSWR. The BPNN inversion model is designed with five input nodes, 4 hide nodes and 1 output node. In order to verify the effective and accuracy of the proposed method, simulation and field data inversion experiments are executed, which shows that the proposed method is effective and provides an efficient solution for the real time ocean surface dynamic parameters' inversion.