Experiments with monocular visual tracking and environment modeling

The purpose has been to find solutions for reliable real-time monocular visual tracking and simultaneous environment modeling. The goal is to estimate the relative motion of a camera with respect to a rigid scene by tracking features such as corners and lines. In the beginning, the 3-D locations of the features are not known accurately, but during the tracking process the model errors are reduced through the integration of new observations. The focus is on modeling measurement uncertainties and selecting the features to be extracted from image frames. Experiments have been performed by using a bank of trackers, each of which calculates estimates for location and motion using measurements of a few features at a time