ANALISIS CAMPURAN SPEKTRAL SECARA LINIER UNTUK KAJIAN PERUBAHAN PENUTUP LAHAN PADA TINGKAT SUBPIKSEL Kasus: Kota Tasikmalaya dan Sekitarnya Berdasarkan Citra Landsat Multitemporal Tahun 1994 dan 2003

• Main Authors: , Riki Ridwana, , Drs. Projo Danoedoro, M.Sc., Ph.D
• Format: Theses and Dissertations NonPeerReviewed
• Published: [Yogyakarta] : Universitas Gadjah Mada 2014
• Subjects: ETD
• Online Access: https://repository.ugm.ac.id/133315/; http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=73900
• Institution: Universitas Gadjah Mada

Land cover in urban areas can be detected through terrestrial surveys or using high-resolution imagery. Although the precision was better but limited to the broad scope and scale generated. Therefore, the image of the medium spatial resolution is needed. However, the use of medium resolution imagery such as Landsat constrained by the presence of mixed pixels uncaused land cover in urban areas tend to be diverse. Pixel mixture will be a source of error in the multispectral classification process, so it takes the analysis to the cedar subpixels that can be obtained through both Linear Spectral Mixture Analysis. One pixel in the Landsat image in this study were separated into four endmember, ie vegetation, impervious surface, open land, and water. The fourth endmember is assumed to represent a linear combination of land cover found in urban areas in the form of proportions in each pixel. This study aims to assess the accuracy of methods are Linear Spectral Mixture Analysis in detecting urban land cover and land cover change detection in Tasikmalaya and surrounding areas from multitemporal Landsat imagery. Accuracy test conducted in this study is divided into two stages, using field measurement data and interviews for image analysis results in 1994, while the accuracy of the test results of the analysis for the year 2003 using DigitalGlobe imagery. Linear Spectral Mixture Analysis in a greater percentage of land cover in each pixel is indicated by the fraction of each endmember image and the image of the RMS error. The results obtained show that the endmember can be separated well, where the RMS error is 0.013 1994 with 94 accuracy rate, 44%, and the RMS error is 0.021 2003 with 97.81% accuracy rate. Of land cover change detection results Spectral Mixture Analysis in Linear Mixed able to demonstrate the changes that occur at each image pixel measuring 30 x 30 meters through the addition or penguranagan endmember proportions in each fraction. Land vegetation cover reduced the extent of 0.81% of the total area, land cover and broader permkaan watertight 0.63%, closing the open land expanded 33.23%, reduced the extent of land cover 18.25% water.