Land Cover Change Detection Analysis for Landslide Monitoring Using SPOT-5 Satellite Images

Landslides have become a common disaster in Malaysia, the incident occurred according to human activity in economic development. Deforestation, urbanization, development of hill slopes and other human land use have given rise to increased risks of landslides. This paper reports the monitoring of the landslide by applying Land Cover (LC) change detection by utilizing two multispectral satellite imagery SPOT 5 for the year 2005 and 2014. ArcMap 10.2.2 software was used in the image processing using Maximum likelihood classification. This study is conducted in Ulu Kelang, Selangor which is located at hilly and sloppy terrain surround by forestry. The results show the method of LC changes and factor of the landslide can be justified via remote sensing image within 10 years’ time. The accuracy of image classification was showed an overall accuracy of 95% (2005) and 93% (2014), while kappa coefficient indicates the method of classification and the images used were excellent quality. During this study period, an urban area has shown an increase of 14.143%, while the green area had decreased by 8.432% due to an increase of interest in urbanization development. There were twelve landslides recorded during the years 2006 to 2014, about 91.66% landslide events were affected build up area which were residential area and road. This analysis showed that human influence or activity could be contributing to landslide occurrences. Therefore, the urbanization development is very important to have proper planning to reduce the susceptibility of landslide occurrences in Ulu Kelang, Selangor.