Study Of Mangrove Coverage Change Based On Object (Obia) Using Satellite Imagery In Dompak Island Province Of Kepulauan Riau

The threats on mangrove forest, either naturally such as climate change or human activities such as landfill, land-use change, and deforestation, can increase the vulnerability of this ecosystem itself Remote sensing is an effective method to use as mangrove monitoring activity because it can be done periodically and can reach a large area. This research aims to analyse mangrove coverage changes in Dompak Island, Kepulauan Riau Province. The method that was used is satellite imagery classification based on object (OBIA) with support vector machine (SVM) algorithm. Satellite imagery data that was used are SPOT 4 in 2007 and Sentinel 2B in 2018 with spatial resolution of 10 x 10 m. Ground check was conducted on September-October 2018 using random sampling method. The classification results of OBIA with SVM algorithm showed 89% accuracy level, 0.86 kappa values with optimum segmentation value of 3. Based on coverage land analysis, there was degradation of 34.19% mangrove area, or about 46.61 ha, since 2007 to 2018.