<![CDATA[Penelitian ini dilaksanakan di Kabupaten Buleleng menggunakan citra Landsat 8 OLI/TIRS (Operational Land Imager/ Thermal Infrared Sensor), dengan tujuan untuk (1) mendeskripsikan metode supervised classification terhadap klasifikasi penutup lahan, (2) mendeskripsikan metode unsupervised classification terhadap klasifikasi penutup lahan, dan (3) membandingkan tingkat akurasi metode supervised classification dengan unsupervised classification terhadap klasifikasi penutup lahan. Metode yang digunakan yaitu metode komparatif dengan membandingkan metode supervised classification dengan unsupervised classification terhadap penutup lahan di Kabupaten Buleleng. Hasil penelitian menunjukkan bahwa (1) diperoleh delapan kelas penutup lahan pada metode supervised classification yang ditentukan oleh pengambilan training area, (2) diperoleh delapan kelas penutup lahan pada metode unsupervised classification yang ditentukan dengan memberikan nilai range statistik, dan (3) tingkat akurasi yang tertinggi dimiliki oleh metode supervised classification yaitu maximum likelihood dengan nilai overall accuracy sebesar 92% dibandingkan dengan metode unsupervised classification (k-means dan ISODATA) yang memiliki nilai overall accuracy yaitu 82,07%. Kesimpulannya adalah untuk deteksi klasifikasi penutup lahan metode yang paling baik dilakukan di Kabupaten Buleleng yaitu supervised classification dengan metode maximum likelihood.This study was conducted in Buleleng Regency using Landsat 8 OLI/TIRS imagery (Operational Land Imager/ Thermal Infrared Sensor), with the aim of (1) describing the supervised classification method for land cover classification, (2) describe the method of unsupervised classification on the classification of land cover, and (3) compare the level of accuracy of the supervised classification method and unsupervised classification on the classification of land cover. The method used is a comparative method by comparing the supervised classification method with unsupervised classification of land cover in Buleleng Regency. The results showed that (1) eight land cover classes were obtained in the supervised classification method determined by the taking of the training area, (2) eight land cover classes were obtained in the unsupervised classification method determined by providing statistical range values, and (3) the accuracy level the highest is owned by the supervised classification method, namely maximum likelihood with the overall accuracy value of 92% compared to the unsupervised classification method (k-means and ISODATA) which has the overall accuracy value of 82,07%. The conclusion is that the detection of land cover classification method that is best done in Buleleng Regency is the supervised classification with the maximum likelihood method.]]>
