Effect of image acquisition and processing parameters on the estimation of crowd-induced dynamic loading on stadium grandstands

In recent years, an increasing number of studies have been developed to assess the dynamic response of structures induced by human loading. Among them, stadiums or sport arenas raise the major interest because of the presence of active people during matches and concerts, leading to high levels of vibrations on the grandstands. To deal with these problems and to guarantee safety during events, the prediction of the force generated by the motion of a large crowd becomes a relevant task. Thanks to its ability to provide full-field displacements even in case of low-resolution images and crowded areas, many studies demonstrate the effectiveness Digital Image Correlation (DIC) for the estimation of crowd motion and induced loading. In this work, a detailed study of the most influencing image acquisition and processing parameters on the accuracy of DIC measurements is presented, in case of images acquired during real events like concerts or football matches. Some values are recommended to be used for future works involving this type of applications.