Method and apparatus for automatically extracting geospatial features from multispectral imagery suitable for fast and robust extraction of landmarks

Abstract

A system for automatically extracting geospatial features from multi-spectral imagery, suitable for fast and robust extraction of landmarks, is presented. The system comprises a computer system including a processor, a memory coupled with the processor, an input coupled with the processor for receiving imagery and user-provided geospatial features, and an output coupled with the processor for outputting the extracted landmarks. The computer system includes a region-growing system, an appending disjointed regions system, and an automatic reseeding system, which are configured to accurately and efficiently extract all regions within the input image that closely resemble the desired geospatial feature. The region-growing system is based on a level set technique, which allows for the fast and accurate extraction of a region evolved from a single user-provided seed point. By automatically reseeding the system when a growing region stops evolving, the system extracts all the desirable regions with minimum user intervention.