Signal Processing and Imaging

Following rapid technological advances, video cameras now offer the ability to capture high-density, time-varying image streams at relatively low cost.

In this project, in collaboration with the I4S team, we aim to estimate the motion of the bar from variations in pixel intensity at the boundaries of the structure (or from pixels corresponding to targets placed on the structure).

Motion estimation relies on specific image processing techniques, the best-known of which are based on Digital Image Correlation (DIC).

Ultimately, certain pixels in the image could serve as sensors for Structural Health Monitoring (SHM), potentially replacing traditional sensors such as accelerometers.

This approach is being tested in the laboratory on a vertical cantilever beam excited at its base by white noise using a shaking table. The displacements estimated through image processing are compared with measurements obtained from laser displacement sensors. In addition, the beam’s natural modes and resonance frequencies are identified using periodogram analysis.

The I4S team will then be able to perform a more detailed modal analysis using signal processing techniques.