Visual vibration analysis

Abstract

This dissertation shows how regular cameras can be used to record and analyze the vibrations of visible objects. Through careful temporal analysis, we relate subtle changes in video to the vibrations of recorded surfaces, and use that information to reason about the physical properties of objects and the forces that drive their motion. We explore several applications of our approach to extracting vibrations from video - using it to recover sound from distant surfaces, estimate the physical properties of visible objects, and even predict how objects will respond to new, previously unseen forces. Our work impacts a variety of fields, ranging from computer vision, to long-distance structural health monitoring and nondestructive testing, surveillance, and even visual effects for film. By imaging the vibrations of objects, we offer cameras as low-cost vibration sensors with dramatically higher spatial resolution than the devices traditionally used in engineering. In doing so, we turn every camera into a powerful tool for vibration analysis, and provide an exciting new way to image the world.