Water Mass Division Algorithm Based On K-means Clustering And Visualization

Water mass analysis, the primary tasks of which are the boundary recognition of water mass and the water mass division, is always an important research area of the oceanographic. By now, the mostly kinds of analysis methods are based on the diagram of temperature and salt. The disadvantage of present water mass analysis methods is that the result of analysis often destroys the space information of water mass and is hard to be communicated with researchers.Visualization in Scientific Computing, which is a kind of newly developed subject in the 1980s, provides widely foreground of water analysis. It is based on computer graphics, data dining and image manipulation technology, and the data generated in the procedure and the calculation results will be displayed on the screen for interactive processing. The researchers can observe and analyze the oceanographic data to discover new rules and relationships of water mass with the intuitionist and third dimension way.Because the character of oceanographic data is huge magnitude, complex structure and many impact factors, it is hard to balance the efficiency of division and interactive visualization. Currently, it is the hotspot of water mass research that how we can effectively divide the water mass quickly and show the results with interactive way.Comparing with the current water mass analysis methods, this paper supplies an algorithm of water mass division based on K-means cluster. It combines the similar conception of water mass and cluster analysis, describes some important factors of water mass division rationally for 3D regular grid data, and applies the K-means cluster into water division for distilling the information of water mass distribution.In the following water mass visualization, this paper researches kinds of volume rendering technology and select ray-casting as experiment based on the data from water mass division. It gets the effective water mass division and 3D visualization of division results by combing the 3D space information with the oceanographic information, such as temperature, salt and water mass boundary.The method for visualization of water mass analysis proposed in this paper can get over the lack of space information in traditional water mass division methods. It makes sure that the result of water mass division reveals the distribution and relationship in real ocean data. The experiment results show that it is corrective and effective by processing the oceanographic data of the neighbor region of Yellow Ocean and East China Sea.