CHARACTERISTICS OF MANGROVE LAND COVER SPECTRAL USING SENTINEL-2A SATTELITE IMAGE IN KEMBUNG RIVER BENGKALIS REGENCY RIAU PROVINCE

Abstract

In visual way, distinguishing mangrove land cover located in Kembung river Bengkalis is hard to do especially if it is using far remote sensing technology. However by applying spectral analyze method, this problem is possibly fixed. It started by taking area sample (training area) 3x3 pixels as many as 10 areas for each classes of land cover, then the result obtained is analyzed by statistical variability test before continued to DMRT test to measure the ability of separation by using separability. Based on the analysis of variability to the spectral value of sattelite vision Sentinel-2A that F-hit (17,626) higher than F-tab (1,98), it showed the treatment has significant effect. The result of DMRT test on reflectancy value of spectral sattelite vision sentinel-2A can be classified into five spectral groups and DMRT test to the mangrove land cover consisted of four groups. Meanwhile, based on separability test there are 36 class pairs, 32 (90,9%) can be separated each way, while 4 (9,1%) can not be separated.