High speed imaging and algorithms for non invasive vibrations measurement

Comunicación presentada en EVACES 2011, 4th International Conference on Experimental Vibration Analysis for Civil Engineering Structures, Varenna (Lecco), Italy, October 3-5, 2011.

Registration of displacements in civil structures is of great importance for assessing their performance, safety and integrity. Traditional sensors like accelerometers have a good response in a wide frequency band. Nevertheless their response depends on the squared frequency so they are more sensitive to high frequencies. This effect can mask vibration in the fundamental modes that may produce larger displacements. In addition to this, double integration from acceleration to shifts is not obvious since many inertial movements are ignored. High speed cameras can provide good results in measuring displacements. Unfortunately, resolution of these cameras is usually lower as higher is the acquiring speed, thus impeding their use for monitoring small vibrations or displacements from long distances. In order to overcome this limitation, subpixel techniques can be applied. These techniques are based on target detection and features extraction of an image, and permit increasing the accuracy by more than one order of magnitude. We have taken an elliptical object as a target. By image detection we can obtain its geometrical parameters and this way we can track rotations around the three main axis and translations in the plane perpendicular to the camera axis. The proposal has been tested both numerically and in in the lab, obtaining accuracies of 0,08 px. The method has been also tested on a real steel column, which is part of an emergency stair. This column was hit by a 1 kg pendulum and the movement was registered with two high speed cameras, an accelerometer and a linear displacement dial. Due to the impact, the dial lost contact with the columns and thus their measurements were no reliable. The accelerometer registered only the high frequency of the signal, while the low frequency was masked and only the two cameras registered the movement in the fundamental frequency. Camera results were validated with a finite element model LS-Dyna and posterior measurements with an accelerometer under a weaker impact, in order to avoid excitement for higher modes.

Spanish Ministry of Science FIS2009-05639-E/FIS; Generalitat Valenciana PROMETEO/2011/21; Universidad de Alicante GRE10-09; Fundación CajaMurcia.