Visualization Of Flow Field Based On Clifford Fourier Transform

Visualization based feature has already been a very attractive and challenging part of the visualization research. Many applications are given in vector field .This paper presents an overview of the techniques and developments of the visualization based feature and studies the key algorithm in the flow visualization. It overcomes the traditional visualization method's disadvantages and can discover important structures and interesting regions of vector field, so can visualize the most important parts of vector field. By this means,irrelative data is removed from the useful data, so data amount is reduced, and sometimes even can get quantitative description of feature. For large number of data, the feature-based method is a necessary choice.The vortex is one of the important features of flow field. This paper detects the vortex flow based on pattern matching method. The vortex flow's feature is detected by computing the greatest similarity of data set's convolution and suitable filter .This method reduces the computation and makes Clifford convolution based on Clifford algebra apply to pattern matching , so it is sensitive to noise. By verification, this method can accurately detect and visualize the vortices of flow field.The convolution and Fourier transform are basic tools for image processing. The fast convolution and characteristic visualization can easily be realized by FFT. Because Fourier transform may accelerate the Clifford convolution, this paper proves fast Clifford Fourier transform is feasible and valid. Fast Clifford Fourier transform can make the Clifford convolution from image space to frequency space, so it can accelerate the convolution computation, cause the computational time reduce greatly.There are a large number of irregular data sets in application. This paper tries to make Non-uniform Fourier Transform extend to Non-uniform Clifford Fourier Transform. Choosing smoothing function as window function and using Matlab to do simulation. The result shows that Non-uniform Clifford Fourier Transform based on smoothing window function is feasible. But if the Non-uniform Clifford Fourier Transform can visualize the irregular flow field effectively or not, much further work is needed.