Detection of structural deformation from 3D point clouds

Abstract

The observation and detection of movements of man-made structures is a noble task in engineering survey, for it is geared towards preservation of life and property. Many different methods exist for detection and monitoring deformations. These methods have served humanity very well over the years. However, most of these methods are point based. Terrestrial laser scanning allows for the monitoring of the whole surface of a structure. All these methods require data to be compared between two or more campaigns. For the data set to be comparable, it needs to be transformed not only into the same coordinate system, but also into the same computational base. An analysis of the stability of reference points by global congruency testing, coupled with the S-transformation enable this to be achieved. In this thesis, the Total station and terrestrial laser scanner were used to detect deformation of a building. The global congruency test was used to detect deformations between two epochs, followed by a single-point analysis which is used in the localization of deformations. After determination of stable scan stations, point cloud data from two epochs was transformed into the same computational base. This enabled point to surface and surface to surface deformations analysis to be undertaken.