| 摘要 |
A system provides a position and orientation (P&O) of a vehicle in a known environment. An environment model is constructed from known parameters. A coarse P&O (90) unit provides approximate vehicle location information of the vehicle to an environment feature extractor (20b) which extracts three-dimensional (3D) identifying characteristics of the environment near the coarse vehicle location. A spatial imaging device (11) provides a vision signal to a vision feature extractor (20a) which extracts identifying two-dimensional (2D) features from the vision signal. A feature-based matcher (70) compares the extracted 2D and 3D features to identify an observation P&O in which the 3D model features will appear as the 2D image features. This observation P&O, after correcting for relative camera displacements between the vehicle and spatial imaging device, is used as the P&O of the vehicle, and also used to determine the inherent error of the coarse locating unit. This error is used to adjust the coarse locating unit's readings until another match is obtained between the 2D and 3D features. <IMAGE> |
