Feature extraction and streamline drawing of 3D time-varying flow field based on physical characteristics

Feature-based flow field visualization is a visualization method at a higher level of abstraction. Normally, only a small part of the data area is the region of interest, and these characteristics can be described very concisely, so huge data reduction can be achieved. This paper uses a discrete method to calculate the physical characteristics of the flow field-divergence and curl, extracts the characteristic area according to the set threshold, and arranges the seed points reasonably. Starting from the seed point, including the positive and negative directions, the fourth-order Runge-Kuta numerical integration performs streamline calculations. And the program uses GPU to accelerate large-scale calculations. Experimental results show that this method can effectively selectively visualize the physical characteristics of the discrete flow field by adjusting the parameters while keeping the physical characteristics of the flow field unchanged, and achieve the simplification and compression of the flow field topology The purpose of the description.