Static accuracy analysis of Vicon T40s motion capture cameras arranged externally for motion capture in constrained aquatic environments

Abstract

While the capabilities of land-based motion capture systems in biomechanical applications have been previously reported, the possibility of using motion tracking systems externally to reconstruct markers submerged inside an aquatic environment has been under explored. This study assesses the ability of a motion capture system (Vicon T40s) arranged externally to track a retro-reflective marker inside a glass tank filled with water and without water. The reflective tape used for marker creation in this study was of Safety of Life at Sea (SOLAS) grade as the conventional marker loses its reflective properties when submerged. The overall trueness calculated based on the mean marker distance errors, varied between 0.257 mm and 0.290 mm in different mediums (air, glass and water). The overall precision calculated based on mean standard deviation of mean marker distances at different locations varied between 0.046 mm and 0.360 mm in different mediums. Our results suggest, that there is no significant influence of the presence of water on the overall static accuracy of the marker center distances when markers were made of SOLAS grade reflective tape. Using optical motion tracking systems for evaluating locomotion in aquatic environment can help to better understand the effects of aquatic therapy in clinical rehabilitation, especially in scenarios that involve equipment, such as an underwater treadmill which generally have constrained capture volumes for motion capture.

Introduction

Motion capture (Mo-cap) is one of the most common methodologies used in biomechanical analysis (Muller et al., 2015, Thewlis et al., 2013). While this instrumentation plays a vital role in areas such as clinical gait analysis to improve treatment of injuries and conditions, it can also be used to address other clinical problems, such as treatment of neuromuscular disorders and cerebral palsy (Andriacchi and Alexander, 2000). Recent studies in clinical rehabilitation have demonstrated the potential benefits of aquatic-based therapies in comparison to land-based therapies (Becker, 2009, Denning et al., 2010, Hinman et al., 2007), and the benefits of exercise in an underwater treadmill have been shown (Conners et al., 2018, Conners et al., 2014, Denning et al., 2010). Factors like density, specific gravity, buoyancy, and other physical principles of water contribute to the advantages of performing physical exercises in an aquatic environment (Becker, 2009). However, in order to quantify these effects, it is important to completely understand human locomotion in an aquatic environment.

Mo-cap systems use different methodologies to collect and analyze human locomotion (Richards, 1999). The reliability and validity of data from such Mo-cap systems has continued to be an area of interest among the scientific community (Eichelberger et al., 2016, Kaufman et al., 2016, Miller et al., 2016, Windolf et al., 2008). In the past decade, the assessments of Mo-cap systems for validity have been predominantly performed on systems that utilize the tracking of retro-reflective marker positions in three-dimensional (3D) space (Eichelberger et al., 2016, Windolf et al., 2008). However, the aforementioned studies have been limited to land-based applications. Rehabilitation scenarios, such as aquatic therapy, have progressed towards trying to evaluate patterns of locomotion while underwater (Kwon and Casebolt, 2006, Silvatti et al., 2013).

Conducting Mo-cap in an underwater environment can be a challenge because the default retro-reflective markers provided by the manufacturers for a Mo-cap system lose the retro-reflective properties due to change in medium once they are submerged. As the surface of the default retro-reflective marker is wetted, it loses its retro-reflective properties. This necessitates identifying a potential retro-reflective material that could retain its retro-reflectivity when submerged in water for aquatic applications. Also, it could be argued that Mo-cap systems that are specifically made for aquatic applications could be used in place of a land-based Mo-cap systems to capture aquatic locomotion (Abdul Jabbar et al., 2017, Lauer et al., 2016). However, in aquatic applications specifically with systems like an underwater treadmill that are used in clinical rehabilitation scenarios, the view window is restricted, and the capture volume is small. Under such conditions, it would be challenging to use an underwater system to capture the human locomotion from inside the unit. Bearing these thoughts, the notion behind utilizing Mo-cap setup arranged externally in this study to evaluate the accuracy derives its motivation from scenarios to detect the motion in such environments. Consequently, this study aims to evaluate the static accuracy of a land-based Mo-cap system arranged externally to track retro-reflective marker position in 3D in different mediums (air, glass and water). The present study also ascertains the influences of medium in the reconstruction of the distances that can affect the overall accuracy of the Mo-cap system.