Parallel Topological Simplification Of Planar Vector Fields Based On Physical Feature

Visualization in scientific computing is a new subject with a rapid development once be proposed since 1980s. Now, it has been widely used in many fields such as Medical, Geoloical Exploration, Weather Forecast, Molecular Biology, Computional Fluid Mechanics and Finite Element Analysis etc, and attract more and more attentions.Planar vector field visualization is one of most important parts of scientific visualization. With the dramatic development of science and technology, the traditional methods can not satisfy the requirements of Scientific visualization for the increasingly wide the fields involved , more and more complex objects, the growing amount of data to be processed and demanding of real-time. In this case, the Feature-based visualization emerged. The selective visualization can filter out irrelevant data, greatly reducing the volume of information processed, and even obtain a quantitative description of the characteristics. The cost advantage of parallel performance and the application requirements promote the development of parallel computing, parallel processing technologies provide an important platform for real-time visualization of massive data. The combination provides a new way to solve real-time visualization of large-scale flow field.At present, the study of playnar discrete topology simplication of the flow field is more popular from the view of Micro and Geometric. Firstly, extract the critical points, and then analysis the nature of them, finally, determine the location of seed points through the represent the geometric characteristics of region. This method is very effective in the core area, but the entire features of flow field performance are not very satisfactory. Fluid movement is strictly dominated by the physical laws, and the geometric characteristics of the flow field are the external performance of the physical characteristics. How to effectively extract the physical charateristics is the key to the scientific visualization of flow field. In this paper, the concept of curl and divergence filter were proposed which use the inegral approach to the physical characteristics of the flow field calculation, after the use of image segmentation and fusion of synthetic images taken from the parameters of physical characteristics extract the parammeters of the flow field, then realize the selective visualization of characteristic areas. Since the physical characteristics analysis and the calculation of stream lines make a large volume of data processing, so parallel processing mechanism was introduced to improve algorithm eficiency, and to meet the needs of remote visualization. Experimental results show that the algorithm can effectively realize the selective visual performance of planar discrete flow field by maintaining the physical characteristics unchanged and adjusting the paramerters of selection, and then describe the overall topology structure quickly and accurately from the physical and Macro point of view, get to the purpose of topological simplification and description compression, and provide an important new approach to deal with large-scale planar flow field.