The Extraction of Edges from Photographs of Quadric Bodies: Part I.
Abstract
For a variety of applications, it is desirable to have a means of reconstructing solid objects from sets of photographs taken from different vantage points. For mathematically well-defined objects, like objects bounded by quadric surfaces, the reconstruction can start with the generation of a two-dimensional line structure representing the projections of edges visible in a photograph. Several line structures then can be used to generate a three-dimensional line structure representing the solid objects. The extraction of a 2-D line structure from a photograph presents two problem areas; the lines to be extracted have to be defined, and a scheme has to be found for recognizing the lines. An outline is given here of an approach that makes use of all available a priori knowledge about the photographs to be processed. In a first step, described in this report, the quantized photograph is subject to a nonlinear transformation (similar to an operator by Hueckel) to generate a 'stroke image' which has fewer and more informative entries than the image array. The performance of the transformation is tested on both artificial and real images. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 01, 1973
- Accession Number
- AD0762589
Entities
People
- E. Urs Ramer
Organizations
- New York University