Machine vision methods for monitoring breakwater armour structures in the model hall environment

M. Tech. Electrical Engineering.This dissertation presents vision-based systems for monitoring model breakwater armour structures. These model breakwater armour structures are subjected to wave simulations in model halls. The goal of this research was to produce a system that is able to detect changes in the model breakwater armour structure. These changes are detected autonomously or semi-autonomously depending on the method being used. The proposed systems are intended to replace the current flicker technique method, which is subject to human error. This dissertation reviews common image processing methods for monitoring changes in three dimensions. This review refines the search to two techniques, namely the stereopsis and fiducial methods. Each method approaches the problem differently. The stereo method treats the entire structure as a holistic 2.5D volume and shows changes in this volume. In the fiducial method special fiducial targets are attached to key areas on the breakwater armour units. The targets are tracked to compute displacement vectors