Bubble tracking in image sequences

Abstract In this paper, we develop a system which can track bubbles in image sequences based on a manually selected initial contour of a bubble. The contours of bubbles are assumed to have circular shapes. Radial scans are used to characterize the position of each bubble. Based on specially selected features, contours of bubbles in subsequent images can be automatically tracked. This novel method reveals to be rather robust and produces good results also under difficult illumination conditions or with reflections on the surface of bubbles. Furthermore, based on a criterion which evaluates the situation of overlapped bubbles, the system can also deal with the overlapping problem and identify the bubbles correctly. After tracking, some parameters such as the bubble departure diameter, bubble flow speed, and bubble volume can be estimated which are important to investigate the heat transfer in a boiling system.