A new visual tracking method for the analysis and characterization of jet flow

Official URL: http://dx.doi.org/10.1016/j.flowmeasinst.2016.08.005

In this paper, we develop a new method to track the evolution of bubbles or droplets in jet flow. Proposed tracker fuses shape and motion features of the individually detected droplets in 2D shadow images and employs the Bhattacharyya distance to assign the closest one among candidate droplet regions. Distinct from the existing droplet tracking techniques in the literature, shapes of the target droplets were not assumed to be circles or ellipses. Instead evolving droplet contours were extracted and analyzed. Proposed tracking algorithm could achieve real time performance with 16 fps in MATLAB environment. Single, double and triple droplets were tracked with the average accuracy of 87%, 87% and 83%, respectively. Experimental results were then evaluated to explain the underlying jet phenomena.