Photogrammetric and Image-Processing Aspects In Qantitative Flow Visualization

Abstract

The development of a measurement system for the visualization, topological classification, and quantitative analysis of complex flows in large-scale wind tunnel experiments is described. A new approach was sought in which the topological features of the flow (e.g., stream lines, separation and reattachment regions, stagnation points, and vortex lines) were extracted directly and preferably visualized in real-time in a virtual wind tunnel environment. The system was based on a stereo arrangement of two CCD cameras. A frame rate of 120 fps allowed measurements at high flow velocities. The paper focuses on the problem of fast and accurate reconstruction of path lines of helium filled soap bubbles in three dimensions (3D). A series of simple algorithmic steps was employed to ensure fast data processing. These included fast image segmentation, a spline approximation of the path lines, a camera model, point correspondence building, calculation of path line points in 3D and creation of a three-dimensional spline representation. The path lines, which contained both velocity and topological information, were analyzed to extract the relevant information.