Structural System Identification from Video Using Finite-Element Shape Functions

This paper presents a novel method for determining structural displacements from video footage using an inverse finite-element (FE) approach. The proposed technique leverages the pinhole camera model to map image pixels to unique spatial coordinates, enabling the calculation of deflection fields from frame sequences. By parameterizing the deflection field using FE shape functions, the method estimates nodal displacements by minimizing the difference between frames of a real video and frames reconstructed using predicted shape functions through an iterative optimization process. Additionally, a particle swarm optimization (PSO) algorithm is employed to estimate the Young’s modulus by minimizing the error between predicted and true structural responses based on the recovered FE state vector history. A synthetic experiment involving a vibrating beam demonstrates the effectiveness of this method. The estimated Young’s modulus of the beam was found to be in excellent agreement with the ground-truth value, validating the method’s potential for video-based structural system identification.