PERBANDINGAN HASIL KLASIFIKASI LIMBAH LUMPUR ASAM DENGAN METODE SPECTRAL ANGLE MAPPER DAN SPECTRAL MIXTURE ANALYSIS BERDASARKAN CITRA LANDSAT - 8

The utilization of remote sensing data is an alternative way that could be used for rapid detection of large coverage hazardous waste area. This study aims to classify the acid sludge contaminated area using Landsat 8 by applying Spectral Angle Mapper (SAM) classification method with two spectral reference sources, namely field spectral measurement using a spectrometer and endmember spectral from the image, and then compare the classification results. The accuracy level of SAM classification result showed that classification using endmember spectral from the image as the reference spectral reached 66,7%, whereas classification using field spectral measurement as spectral reference only reached 33,3%. The accuracy level of Spectral Mixture Analysis (SMA) classification result showed that classification using endmember spectral from the image as the reference spectral reached 62,5%.  The affecting factors for the low accuracy is the significant differences of the spectral profiles obtained from spectrometer with spectral Landsat-8 due to differences of spatial and altitude Abstrak Pemanfaatan data penginderaan jauh merupakan salah satu alternatif yang dapat digunakan untuk deteksi daerah tercemar limbah B3 secara cepat dengan wilayah yang luas. Penelitian ini bertujuan untuk mengklasifikasi daerah tercemar lumpur asam menggunakan data Landsat 8 dengan metode Spectral Angle Mapper (SAM), kemudian membandingkan hasil klasifikasiSAM menggunakan spektral referensi berdasarkan pengukuran spektrometer dengan spektral yang diperoleh dari endmember citra. Tingkat akurasi klasifikasi SAM dengan spektral referensi berdasarkan endmember citra adalah sebesar 66,7 %, sedangkan dengan menggunakan referensi spektrometer hanya mencapai 33,3 %. Tingkat akurasi klasifikasi Spectral Mixture Analysis (SMA) dengan spektral referensi berdasarkan endmember citra adalah sebesar 62,5 %. Faktor yang mempengaruhi rendahnya akurasi adalah perbedaan yang signifikan antara profil spektral yang diperoleh dari spektrometer dengan spektral Landsat-8 akibat perbedaan spasial dan ketinggian. The utilization of remote sensing data is an alternative way that could be used for rapid detection of large coverage hazardous waste area. This study aims to classify the acid sludge contaminated area using Landsat 8 by applying Spectral Angle Mapper (SAM) classification method with two spectral reference sources, namely field spectral measurement using a spectrometer and endmember spectral from the image, and then compare the classification results. The accuracy level of SAM classification result showed that classification using endmember spectral from the image as the reference spectral reached 66,7%, whereas classification using field spectral measurement as spectral reference only reached 33,3%. The accuracy level of Spectral Mixture Analysis (SMA) classification result showed that classification using endmember spectral from the image as the reference spectral reached 62,5%.  The affecting factors for the low accuracy is the significant differences of the spectral profiles obtained from spectrometer with spectral Landsat-8 due to differences of spatial and altitude