Smartphone based parameter estimates of a dynamic oscillator using high-speed video imaging and incremental discriminating colour learning

Abstract

Image-based systems are increasingly being used for Structural Health Monitoring (SHM) applications. Video-based motion tracking algorithms can be used to analyse dynamic responses characterised by low frequencies, large deflections and low damping ratios. The advantages of image processing over other methods include the ability to track multiple points on a structure, its scalability, and its ease of use. Standard video acquisition devices are limited in their ability to assess dynamic responses and identify natural frequencies or damping ratios of structures due to the relatively low sampling rate, or frame rate. As such, there becomes a need to use video cameras that possess the ability to record at high frame rates — a feature that is becoming increasingly common on modern smartphones. This paper demonstrates how such video cameras can be used to estimate natural frequencies and viscous damping ratios of structures by considering a Single Degree of Freedom (SDOF) linear system undergoing free vibrations. The slow-motion feature on a Nexus 6P Smartphone was used to capture the dynamic response of the vibrating system. The video was assessed by an Incremental Discriminative Colour Tracking (IDCT) algorithm which tracked the position of points on the system, from which the natural frequency and damping ratio could then be extracted. The results were compared to a reference accelerometer and theoretical estimates. This paper acts as an evidence base for the evolving capabilities of smartphone based monitoring, and ultimately, shows that smartphones have value as a tool for the cost-effective assessment of structures.