Feature extraction of complex ocean flow field using the helmholtz-hodge decomposition

Flow visualization is an important research field in scientific visualization and feature detection is the one of the core problems. This paper presents a novel approach for complex ocean flow visualization and analysis by applying the helmholtz-hodge decomposition theory to the feature extraction problem. We give an efficient implementation on regular grids by solving a large-scale sparse group of linear equations. And to accelerate the computation process, we have used the GMRES parallel library. By making full use of the anti-noise property of the decomposition results, well-designed algorithms are used to identify the features such as critical points and vortices. To illustrate the ability of our techniques, experiments on realistic datasets are conducted. Experimental results demonstrate that our methods are helpful to deeply understand the flow field.