Pemetaan Mangrove di Pulau Lancang Menggunakan Drone dan Sentinel-2A dengan Metode Object Based Image Analysis dan Pixel Based Analysis

Abstract

Mangrove forests in Indonesia has decreased by 30% to 50% of their total area in the last 50 years. Therefore, it was necessary to monitor mangroves in Indonesia. Percentage of mangrove canopy cover obtained with hemispherical photography method. The percentage of mangrove canopy cover on Lancang Island resulted in an average value of 63.12%, taken from 90 field points. The highest percentage value of mangrove canopy cover was 80.64%, and the lowest value was 30.76%. The object-based classification was done by a Support Vector Machine (SVM) algorithm. The pixel-based classification was done by Maximum Likelihood (MLH) algorithm. The accuracy test results of object based classification on drone and Sentinel-2A images are 96.38% and 92.77%. The accuracy test result of pixel-based classification on Sentinel-2A is 87.95%. The accuracy test result stated that object-based classification method and pixel based classification method give a good results.