Empirical mode decomposition of hyperspectral images for segmentation of seagrass coverage

Seagrasses are an integral part of the marine ecosystem, and can provide information about their environment based on their surface content. In particular, epiphytes and epifauna on seagrass blades are of interest to scientists. Empirical mode decomposition is applied to hyperspectral images obtaine...

Full description

Saved in:
Bibliographic Details
Main Authors: Mehrubeoglu, Mehrube, Trombley, Chris, Shanks, Susan E., Cammarata, Kirk, Simons, James, Zimba, Paul V., McLauchlan, Lifford
Format: Conference Proceeding
Language:English
Subjects:
Biology
Blades
classification
Empirical mode decomposition
hyperspectral image processing
Hyperspectral imaging
Image segmentation
Noise measurement
seagrasses
segmentation
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Seagrasses are an integral part of the marine ecosystem, and can provide information about their environment based on their surface content. In particular, epiphytes and epifauna on seagrass blades are of interest to scientists. Empirical mode decomposition is applied to hyperspectral images obtained from seagrasses to separate hyperspectral data into component modes, and then to segment and classify the seagrass coverage. A sample spectrum is taken from the image for reference for each of the classes (seagrass leaf, tubeworm, epiphyte). Hypothesis testing on the higher modes for an entire image gives a semi-automated algorithm for classifying the contents of unknown spectra. A classifier is developed to segment the seagrass hyperspectral images and identify epiphytes on the seagrasses.
ISSN:1558-2809
2832-4242
DOI:10.1109/IST.2014.6958441