Land cover changes mapping in cameron highlands using high resolution satellite and unmanned aerial vehicle imageries

Abstract

Agriculture and tourism are two important economic activities in the hilly area of Cameron Highlands, Pahang, Malaysia. Land opening for agriculture and construction of settlements and hotels to cater for tourism activities are rapidly and continuously ongoing in this area. However, improper planning of these activities has resulted in various environmental issues such as landslide hazards. This research is undertaken to assess the land use and land cover (LULC) changes occurred in the study area for a period of 12 years (2001-2013) using high resolution optical satellite images (IKONOS and QuickBird) and unmanned aerial vehicle (UAV) images from a fixed wing Helang. An object based classification technique was used to classify the satellite images and UAV images into seven LULC classes, namely, forest, agriculture, grass, bare land, urban, water body and areas affected by landslides. The results obtained from the classification technique were verified using land use maps of 2003, 2008 and 2015 that were obtained from the Department of Town and Rural Planning. The overall accuracy and Kappa Coefficient values (values in brackets) of the LULC classification are 86.67% (0.84), 83.89% (0.81), and 93.80% (0.93) for 2001, 2007 and 2013 respectively. Post classification change detection technique was applied in this study to identify LULC changes. Results of the classification show that the forest area decreased consistently from 2001 (196.08ha) to 2007 (180.73ha) and to 2013 (160.09ha). On the other hand, the built-up area, increased during the years from 47.77ha in 2001 to 58.25ha in 2007 and to 63.43ha in 2013. In these periods, a slight increase was noticed in the agriculture and grass lands, however, water bodies did not change much. In general, bare soil areas have only minor changes. Areas affected by landslides are detected in the UAV image and it covered an area of 3.66ha. In conclusion, this study show that the optical satellites and UAV images can be processed to produce accurate classification map, therefore useful for the local authorities to identify land cover changes, furthermore to monitor land encroachment activities and to reduce landslide hazards from occurring and to mitigate its effect.