Analysis of land covers over Northern Peninsular Malaysia by using ALOS-PALSAR data based on frequency-based contextual and neural network classification technique

Hwee San Lim(1), Mohd Zubir Mat Jafri(1), Khiruddin Abdullah(1) and Nasirun Mohd. Saleh(1)

(1) Universiti Sains Malaysia, School of Physics,, 11800 Penang, Malaysia


Optical and microwave remote sensing data have been widely used in land cover and land use classification. Optical satellite remote sensing methods are more appropriate but require cloud-free conditions for data to be useful especially at Equatorial region. In Equatorial region cloud free acquisitions can be rare reducing these sensors' applicability to such studies. ALOS-PALSAR data can be acquired day and night irrespective of weather conditions. This paper presents a comparison between frequency-based contextual and neural network classification technique by using ALOS-PALSAR data for land cover assessment in Northern Peninsular Malaysia. The ALOS-PALSAR data acquired on 10 November 2006 were converted to vegetation, soil, urban and other land features. The ALOS-PALSAR images of the study area were chosen for land cover mapping using the standard supervised classification techniques (maximum likelihood, minimum distance-to-mean and parallelepiped). The PALSAR data of training areas were choose and selected based on the high resolution optical satellite imagery and were classified using supervised classification methods. Supervised classification techniques were used in the classification analysis. The best supervised classifier was chosen based on the highest overall accuracy and Kappa statistic. Based on the result produced by this study, it can be pointed out the utility of ALOS-PALSAR data as an alternative data source for land cover classification in the Peninsular Malaysia.



  Higher level                 Last modified: 07.05.06