Integration of multiple vision systems and toolbox development

Depending on the required coverage, multiple cameras with different fields of view, positions and orientations can be employed to form a motion tracking system. Correctly and efficiently designing and setting up a multi-camera vision system presents a technical challenge. This thesis describes the development and application of a toolbox that can help the user to design a multi-camera vision system. Using the parameters of cameras, including their positions and orientations, the toolbox can calculate the volume covered by the system and generate its visualization for a given tracking area. The cameras can be repositioned and reoriented using toolbox to generate the visualization of the volume covered. Finally, this thesis describes how to practically implement and achieve a proper multi-camera setup. This thesis describes the integration of multiple cameras for vision system development based on Svoboda\u27s and Horn\u27s algorithms. Also, Dijkstra\u27s algorithm is implemented to estimate the tracking error between the master vision system and any of the slave vision systems. The toolbox is evaluated by comparing the calculated and actual covered volumes of a multi-camera system. The toolbox also is evaluated for its error estimation. The multi-camera vision system design is implemented using the developed toolbox for a virtual fastening operation of an aircraft fuselage in a computer-automated virtual environment (CAVE) --Abstract, page iii