Object-oriented classification of rubber plantations from Landsat satellite imagery

Abstract

Due to increasing global demand for natural rubber products, rubber (Hevea brasiliensis) plantation expansion has occurred in many regions where it was originally considered unsuitable. However, accurate maps of rubber plantations are not available, which substantially constrain our understanding of the environmental and socio-economic impacts of rubber plantation expansion. In this study, the rubber plantations was accurate mapped from Landsat satellite imagery based on object-oriented classification method in Yangjiang State Farm in Hainan Island in 2010. The results show that: (1) The rubber plantation area in Yangjiang State Farm was estimated at 5866 hm2 in 2010, which was slightly higher than the stand inventory data (5190 hm2) in 2009. (2) The resulting rubber plantation map has a high accuracy according to the confusion matrix by using the ground truth ROIs. The overall accuracy is 90% and the kappa coefficient is 0.9. It showed that object-oriented classification method is suitable for mapping rubber plantation from Landsat satellite imagery.