A feature-based seeding method for multi-level flow visualization

Streamline is one of the most important methods for flow visualization and selecting seeds plays a key role for generating streamlines. So we propose an improved algorithm which combines information entropy and critical points. We use information entropy to find the subfields with drastic changes, and then employ critical points to identify the topology features to control the details of the subfields. In this way, not only the areas of dramatic changes in the flow field can be covered, but also the topology features in the field can be captured. In addition, we design a series of interactive operation to display the field distinctly in both the whole field and the subfield. For the whole field, the method we proposed is used to seed which can display as much information of the field as possible macroscopically. For the subfields, based on our interactive design, we place seeds only on the subfields which are selected by users. This method can not only use moderate streamlines to display overall field without visual clutter but also make sure that there are enough streamlines depicting the local details of the field when amplified to a local region. Finally the experimental results show the effectiveness of the algorithm. © 2016, Institute of Computing Technology. All right reserved.