Data fusion between strain data from strain gauges and displacement data from image registration

In structural analysis, different tools can be used to measure the position, distance, or vibration of an object. Such instruments can be divided into two basic categories: contact and non-contact instruments. Among the types of non-contact measurements, there are image registration and optical flow techniques. With the use of a camera, it is possible to calculate the real-time movement of all the points of the structure taken up by the camera. With these techniques, while on the one hand, it is possible to obtain the displacement in numerous points of the structure without the need for complex instrumentation, on the other hand, the data obtained are affected by a high level of noise. Among the types of contact measurements, there are strain gauges, which provide the strain values of the structure at the points where they are installed. With strain gauges, using shape-sensing techniques, it is possible to calculate the deformation of the structure in real-time. But, if on the one hand, the data provided by the strain sensors are very precise, on the other it is necessary to install a large number of them in order to reconstruct the deformation of the structure with a certain accuracy. The aim of this thesis is to perform a data fusion between photogrammetric and strain data to improve the displacement field that would be obtained using the two techniques independently. The data fusion will be performed on two experimental tests regarding two different beams, using a video camera and strain gauges. In this thesis, image registration and shape-sensing techniques will first be illustrated. After a brief review of the instruments used for testing, the descriptions of the experimental tests and of the algorithm developed to perform the data fusion will follow. Finally, the results of the analysis will be presented, with an evaluation of the effectiveness of the method together with the conclusions.