Edge extraction of radar and optical images
Abstract
Examination of radar and optical images indicate the most distinguishable common features of the two types of images are the outlines of the various natural as well as man-made objects. It is expected that edge features can be used to advantages in the recognition and matching of objects using a digital computer. The paper describes an edge extraction procedure which is capable of extracting salient edges from the image and remove many of the edges extracted from the background and shadows around the objects. The procedure involves these three steps: (1) eliminate background edges by making the bright background uniformly bright and the dark background uniformly dark; (2) retain only salient edges by adjusting the edge threshold to minimize the weak edges; and (3) thicken the remaining edges before scene matching to minimize the effect of minor misregistration.
- Publication:
-
Conference on Pattern Recognition and Image Processing
- Pub Date:
- 1979
- Bibcode:
- 1979prip.conf..150W
- Keywords:
-
- Geometrical Optics;
- Image Processing;
- Imaging Techniques;
- Radar Imagery;
- Target Recognition;
- Digital Computers;
- Edges;
- Frequency Distribution;
- Matching;
- Optical Thickness;
- Instrumentation and Photography