Aplikasi dan Evaluasi Citra ALOS PALSAR Resolusi 50 m, Resolusi 12,5 m, dan Resolusi 6 m untuk Identifikasi Tutupan Lahan (studi kasus di Kabupaten Humbang Hasundutan, Kabupaten Tapanuli Utara, dan Kabupaten Samosir)

Abstract

The application of remote sensing technology for rasionalization land resources data usage has more benefit than data collecting in conventional ways. Remote sensing techniques have been used in Indonesia for forest inventory and forest resources monitoring. Indonesia‟s geographical position where is in the tropics area that has vast cloud cover limits the use of optical satellite image data. To overcome the optical satellite image weakness, active remote sensing system (radar) is currently available. One of the satellites carrying radar sensor which was launched by the Japanese government is ALOS (Advanced Land Observing Satellite) with PALSAR (Phased Array type L-band Synthetic Aperture Radar) as its one of the sensor. This sensor is an active microwave sensor which has capability to penetrate thick cloud layer. The objective of this research is to analyze the ability of ALOS PALSAR satellite image of 50 m, 12.5 m, and 6 m resolutions on land cover classification at Humbang Hasundutan Regency, North Tapanuli Regency, and Samosir Regency. The observation has delivered 16 land cover classes, which are as follow: water body, harvested rice field, cultivating rice field, just-planted rice field, bushes, coffee plantation, dry land agriculture, town settlement, village settlement, bare land, nature forest, Eucalyptus plantation year 2010, Eucalyptus plantation year 2009, Eucalyptus plantation year 2006, Eucalyptus plantation year 2007, and Eucalyptus plantation year 2008. The result of digital interpretation of the ALOS PALSAR satellite image 50 m, 12.5 m and 6 m resolutions obtained 6 classes of land cover which are as follows: water body, rice field, dry land agriculture, settlement, bare land, and tree vegetation by different value of proportion correct. ALOS PALSAR of 50 m, 12.5 m and 6 m resolutions, each has proportion correct in consequence 62.1%, 50.5% and 68,9% Visual analysis, the result of ALOS PALSAR image with resolution 50 m obtained 8 classes of land cover,which are as follows: water bodu, dry land agriculuture, settlement, harvested rice field, young forest plantation, middle-old forest plantation, nature forest, dan bare land. Whereas, accuration and separability analysis of ALOS PALSAR image of 50 m resolution has regular accuracy value 59.15% and kappa accuracy 53.48%. The result of the interpretation raising-ability analysis of ALOS PALSAR image of 12.5 m and 6 m resolutions to the ALOS PALSAR image of 50 m resolution are able to provide additional land cover classes, clarify the delineation and make easier the identification of land cover at the satellite image.