Computer Vision Estimation of Physical Parameters and Its Application to Power Requirements of Natural and Artificial Swimmers

Abstract

A useful measure of efficiency of transport in aquatic animals and autonomous underwater vehicles is cost of transport. Often, cost of transport data on specific animals or platforms is not readily available or does not fit specific use cases, but images are readily available. In this work, we present a methodology to synthesize such data without the need for a specimen or laboratory tests. We propose a computer vision in a methodology called Ika-Fit to determine important physical characteristics, such as surface area, slenderness ratio, and mass, that are used for a cost of transport model. The Ika-Fit method provides a good estimation of parameters when compared to biological data and robotic platforms. These parameters are estimated for existing engineered systems, and the model is compared to published data; the model is found to demonstrate higher accuracy using fewer parameters in estimating cost of transport over existing methods.