Analysis of reflectance spectra of tropical seagrass species and their value for mapping using multispectral satellite images

Mapping of the distribution of individual seagrass species is essential for any attempts to manage seagrass ecosystems. It is therefore important to understand how the spectra of different seagrass species vary, in order to establish their unique absorption features and how these can be utilised for...

Full description

Saved in:
Bibliographic Details
Published in:International journal of remote sensing 2019-12, Vol.40 (23), p.8955-8978
Main Authors: Wicaksono, Pramaditya, Fauzan, Muhammad Afif, Kumara, Ignatius Salivian Wisnu, Yogyantoro, Rifka Noviaris, Lazuardi, Wahyu, Zhafarina, Zhafirah
Format: Article
Language:English
Subjects:
Absorption
Accuracy
Algorithms
Classification
Clustering
Divergence
Ecosystem management
Ecosystems
Image analysis
Image classification
Image segmentation
Mapping
Procedures
Reflectance
Remote sensing
Resampling
Satellite imagery
Satellites
Sea grasses
Shape
Spaceborne remote sensing
Species
Species classification
Spectra
Strategic management
Tropical climate
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Mapping of the distribution of individual seagrass species is essential for any attempts to manage seagrass ecosystems. It is therefore important to understand how the spectra of different seagrass species vary, in order to establish their unique absorption features and how these can be utilised for mapping by making use of remote-sensing images. This paper presents measurements of the reflectance spectra between 400 and 900 nm for nine tropical species of seagrass. Continuum removal and multispectral resampling procedures were applied to the spectra. Dendrogram analysis was carried out to identify species clustering as the basis for a mapping scheme. Spectral Angle Mapper (SAM) and Spectral Information Divergence (SID) approaches were employed for the classification of seagrass species using WorldView-2 images and measured spectra as the input endmember. Classification Tree Analysis (CTA) and an image segmentation approach using CTA (Object-Based Image Analysis - OBIA) were performed as a means of comparison. The results indicate that the absorption features and overall shape of the spectra for all seagrass species are relatively similar, and implied that the major differences are attributable to the absolute reflectance values. Consequently, SAM and SID produced results of low accuracy (
ISSN:0143-1161
1366-5901
DOI:10.1080/01431161.2019.1624866