Image processing and analysis of colored particle motions in turbulent flow.

摘要

Image processing and analysis techniques are used to extract quantitative information from visual data obtained in fluid mechanics. Analysis of films obtained from turbulent fluid flow allows the researcher to obtain direct measurements of the instantaneous velocity vector field in both space and time.;Films from a previous work were digitized in full color at image resolution of 2048 x 2048 pixels and pixel resolution of 8 bits per pixel. Techniques were applied to resolve color separation observed at the edges of the particles, remove registration lines recorded on the original films, and suppress low intensity noise. Procedures and algorithms were developed to automatically identify the particles in a full color frame, establish stereopairs, and convert the two sets of x, y coordinates recorded on the left and right scene on the film into the three-dimensional particle coordinates. The predictor-corrector method was formulated in order to automate the tracking process. Oniy three frames have to be fully tracked manually in order to apply the predictor-corrector model. The model accounts for path and velocity coherence. It was also utilized to establish stereopairs by simultaneously tracking the particles in the left and right views. In this way, stereo and motion processes were unified.;The results showed that image analysis based on the predictor-corrector model made it possible to track the particles frame by frame and establish correct stereopairs. Particles in every third frame were computer predicted. By predicting every fourth or so frame, the predictor-corrector technique can reduce computations even further. However, the distance between the first and the last frames used in the model should be less than the turbulence microscale.