Comparison of seagrass cover from multi-scale imagery

Abstract

Abstract Mapping of seagrass bed can be done by using different characteristics of satellite images data such as Landsat-8 and Sentinel-2. However, there has been no research compared to the results from those images in Indonesia. This study aims to analyze and compare spatially the condition of seagrass bed from Landsat-8 and Sentinel-2 imagery covered Kodingareng Lompo island. Spatially sharpened (Pan-sharpening) Landsat-8 imagery from 30 to 15 meters resolution in four spectral bands (visible and NIR) were used in this case. The results showed the four classes seagrass cover percentage consisted of sparse class (0 – 25%), medium-class (26 – 50%), dense class (51 – 75%), and very dense class (76 – 100%). Classification of seagrass cover class and water column correction (Depth Invariant Index) value differed between the two imageries. Accuracy test of seagrass cover classification from Landsat-8 (88.1%) and Sentinel-2 (89.2%) were both suitable to use since they were higher than the standard value (>85%). Differences in the characteristics of both images cause differences in the results of the seagrass cover classification.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint