A technique is described for precise registration of two images of the same area, taken under different conditions. The technique, called AUTO-MATCH, involves digital preprocessing of the images to extract edge contours, followed by a correlation of corresponding edge end points. The availability of an array of endpoints makes possible subpixed registration accuracy. The technique is applied to a system for assessment of geometric image quality, to be installed at NASA-Goddard. This system measures the registration vectors over an array of window pairs from LANDSAT images, so as to determine the distortion between them. Similar measurements could be used to infer relative positioning of the spacecraft. The technique is considered for updating the inertial navigation systems of missiles or aircraft. Scene matching between an image obtained aboard the vehicle and a stored reference can eliminate the drift of an inertial platform in order to demonstrate this capability in real time. A laboratory demonstration of AUTO-MATCH is described.
A digital scene matching technique for geometric image correction and autonomous navigation
1979-04-01
Conference paper
No indication
English
Autonomous Mobile Robot Navigation Using Scene Matching with Local Features
British Library Online Contents | 2016
|An Autonomous Relative Navigation Algorithm Based on Interframe and Scene Matching
Springer Verlag | 2025
|Multi-Region Scene Matching Based Localisation for Autonomous Vision Navigation of UAVs
Online Contents | 2016
|Suitability of scene matching for unmanned vehicle navigation
Tema Archive | 1998
|Differential correction technique to the rotation error in autonomous navigation constellation
British Library Online Contents | 2014
|