Hyperspectral Imaging for Bottom Type Classification and Water Depth Determination
Abstract
Many recreational, military, and commercial activities take place in shallow coastal waters; therefore, interest is high in characterizing these areas. A variety of methods have been employed to determine water depths and classify the bottom using remote sensing. This research proposes to apply Philpot's principal components algorithm for bathymetric mapping to a MISI hyperspectral image, whereas previously this approach has been used on synthetic data. A description of the principal components algorithm is presented along with an outline of how it was applied to airborne hyperspectral images. The algorithm takes advantage of the ability to implement a deep-water correction, and in this linearized space, perform an eigenvector analysis to determine maximum variance in the data, which is related to depth. Unsupervised classification was performed on the first two principal component scores, resulting in a qualitative depth map and bottom type map.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 11, 2000
- Accession Number
- ADA383049
Entities
People
- Nikole L. Wilson
Organizations
- Air Force Institute of Technology