Representation And Recognition Of Elongated Regions In Aerial Images

Charles A. Lipari; Charles A. Harlow; Mohan M. Trivedi

doi:10.1117/12.947020

Translator Disclaimer

29 March 1988 Representation And Recognition Of Elongated Regions In Aerial Images

Charles A. Lipari, Charles A. Harlow, Mohan M. Trivedi

Author Affiliations +

Proceedings Volume 0937, Applications of Artificial Intelligence VI; (1988) https://doi.org/10.1117/12.947020
Event: 1988 Technical Symposium on Optics, Electro-Optics, and Sensors, 1988, Orlando, FL, United States

Abstract

The recent advances in model-based image understanding systems for aerial imagery have used intermediate-level interpretations of the low-level segmentations in terms of generic shape properties. This paper is concerned with those features recognizable in the context of having an elongated shape as evidenced from the planar point sets obtained from spectral classifications. The basic processing technique can be described as linear-strip point clustering, using the minimum spanning tree as a connectivity constraint. The algorithms are applied to classifications of suburban road networks. The major elongated regions are detected for the given strip width, and where connectivity can be established across any gaps in the segmentation.

Citation Download Citation

Charles A. Lipari, Charles A. Harlow, and Mohan M. Trivedi "Representation And Recognition Of Elongated Regions In Aerial Images", Proc. SPIE 0937, Applications of Artificial Intelligence VI, (29 March 1988); https://doi.org/10.1117/12.947020

ACCESS THE FULL ARTICLE