ALOS PALSAR Image for Landcover Classification Using Pulse Coupled Neural Network (PCNN)

Abstract: This research examined the landcover classification using remote sensed image radar system and using a computational vision system of cortex. The aims of this research, 1) to use SAR (Synthetic Aperture Radar) Image for landcover classification using PCNNN, 2) to utilize image processing software for landcover classification using PCNN, 3) to analyze the ability of SAR image for landcover classification using PCNN. Method was used in this research, remote sensed image processing using artificial neural network with PCNN architecture. The field sampling frame work of the landcover classes using stratified random sampling. The field data were analyzed by confusion matrix to determine the level of accuracy. The results of this research are landcover classes of SAR image classification using PCNN, landcover classification using SAR image and PCNN utilized by image processing software, produce three classes of landcover consist of buildings, vegetation, and open land/water body, with an accuracy level of 70,03%.

Keywords: Landcover, Synthetic Aperture Radar, Pulse Coupled Neural Network (PCNN), Remote Sensing