Uji Akurasi Klasifikasi Terbimbing Berbasis Piksel Pada Citra Sentinel 2-A Menggunakan Citra Tegak Resolusi Tinggi Tahun 2019 di Kota Padang

Dedy Fitriawan (Program Studi Geografi,Universitas Tamansiswa Padang)

Jurnal Azimut
Vol 3 No 01 (2020): Volume 3 No. 01

Currently, land cover data sourced from satellite imagery are increasingly being used with advances in sensing technology that are increasingly sophisticated. Maximum likelihood is one of the digital image classification methods that have long been used for pixel-based image classification. This study aims to classify land cover on Sentinel-2A satellite imagery using the maximum likelihood method to see the level of accuracy in that method. The accuracy test is carried out by comparing the results of the classification of the land cover map on the Sentinel-2A image with the sample in the High Resolution Upright Image with the acquisition date which is said to be Worldview and Geoeye imagery in 2019. From the classification results obtained an overall value of 90.81% with the type of land cover highest level of accuracy is wetlands and built-up areas. Meanwhile, the type of cover with the lowest accuracy or the most errors occurred in mixed garden types.