Tracking Bubbles for Outdoor Flow Characterization

Research into Micro Air Vehicles (MAVs) and other small, dynamic robots often involves quantification of the dynamic behavior of these vehicles. While the most interesting behavior occurs in unconstrained environments, the behavior which is easiest to analyze occurs in constrained, laboratory environments. In either case, video analysis is common for many of the same reasons it is commonly used in biological research applications low disturbance to the system, potentially high spatial resolution, and easily adaptable to laboratory scales [1]. Tools like those developed by [1] are often used for image analysis, but these tools have some limitations and are designed for more general objectives. For any given analysis, a great deal of time is often spent with a human operator manually identifying points in each frame of interest. Figure 1 shows typical trajectory and velocity information extracted using this manual method. While crucial to the underlying research in terms of both final results and iterative feedback during development, the high time commitment creates a barrier to its use.