<![CDATA[Mapping of mangrove distribution is important as basic information in mangrove resource management. development of remote sensing technology with multispectral unmanned aerial vehicle (UAV) with high spatial resolution. This study aims to determine the classification and distribution of mangrove genera using a pixel-based classification method and calculate the accuracy level of mangrove genus classification using a multispectral unmanned aerial vehicle (UAV) in Lancang Island Waters, Kepulauan Seribu. This research was carried out in August 2023 by obtaining 481 mangrove genus observation points using the DJI Phantom 4 multispectral drone. Image classification was processed using a pixel-based classification method with two classification levels, including level 1 (mangrove), resulting in an area of 18.72 ha. Level 2 (mangrove genus) uses guided classifications such as Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Random Forest (RF). Based on the classification results, the best results were obtained using the RF algorithm with an accuracy of 89.78% and a kappa index of 0.51, followed by the SVM algorithm with an accuracy of 89.78% and a kappa index of 0.45, then using the KNN algorithm with an accuracy of 88.32% and a kappa index of 0.43.]]>
Copyrights © 2024
