Land cover/use classification by using ALOS-PALSAR and ALOS-AVNIR data

Remote sensing is a very useful method in application of establishing land cover map. It has been widely used for detecting and analyzing land cover/use feature in our environment especially in the developing country like Malaysia. The objective of this study is to investigate the use of quad polarization ALOS-PALSAR data for land cover/use classification. Speckle filtering was an improvement over the unfiltered ALOS-PALSAR data. The use of Mean texture measure was found to be advantageous. A maximum likelihood decision rule is utilized to determine different land-cover classes such as paddy, forest, oil palm, water, and urban. The ALOS-PALSAR data training areas were chosen based on optical satellite imagery. Ground reference data from sites throughout the study area were collected for validation. The land cover information was extracted from the digital data using PCI Geomatica 10.3.2 software package. The high accuracies indicated that the land cover/use can be mapped accurately using ALOS PALSAR data.