Integrating neural networks and correlation for identifying targeted control points

Abstract

The main goal of this work is to present new tools for automated identification
of photographic images of ground targets for aerophotogrametry applications
such as aerotriangulation. The identification of the targeted control points
images was done by using the concepts of statistical correlation, artificial neural
networks and mathematic morphology with emphasis on binary erosion. It was
established procedures for identifying automatically common points in a pair of
photographs. With the purpose of verifying the consistency of the identified
points a fotogrammetric model of the test area was mounted.