Three Dimensional Tracking of Four Point Planar Patterns Using Corner

Abstract

<p>This thesis concerns tracking of planar shapes using a single camera view. Two dimensional silhouette projections of planar shapes randomly oriented and positioned and rotational orientation with respect to the camera. This information can then be provided to robots or automated guided vehicles for automated assembly, docking, shape identification and retrieval.</p> <p>Silhouettes are initially encoded in the system as a linked list of vertex points representing changes in the direction of the contour. This contour is employed to get an initial estimate of the shape's position. Corner feature points of the shape are then more accurately located using a fast corner location algorithm. A photogrammetric solution is then applied to the projection points of the corners to determine the three dimensional position of the shape. The photogrammetric technique employed does not require a least squares fit as does the traditional photogrammetric approach. This new approach is contrasted with the traditional method in terms of accuracy and computational complexity.</p>