Abstract—Rubber tree is an important economics crop in Thailand as well as the Association of South East Asian Nation (ASEAN) region. In addition, it also help to absorb the carbon dioxide stored in the form of biomass. Biomass of plants is one of the essential variables in explaining the climate system and carbon cycle. The objective of this study was to estimation the above ground biomass of the rubber tree using high spatial resolution spaceborne multispectral sensor (i.e., WorldView-2). The 8 spectral bands from WorldView-2 imagery were used as input variables of Stepwise Multiple Linear Regression and Artificial Neural Networks for estimate the biomass of the rubber tree at Paklok sub-district, Thalang district, Phuket Province. The results showed that Artificial Neural Networks provide the most accurate (Root Mean Square Error (RMSE) = 11.97) when compared with stepwise multiple linear regression (RMSE = 13.07). We hope that the methodology presented in this study can be used as a guideline for study in other area and for rubber tree plantation management or predictions the rubber yield in the future.

Index Terms—rubber tree, biomass, remote sensing, stepwise multiple linear regression, neural networks

Cite: Kanjana Yasen and Werapong Koedsin, "Estimating Aboveground Biomass of Rubber Tree Using Remote Sensing in Phuket Province, Thailand," Journal of Medical and Bioengineering, Vol. 4, No. 6, pp. 451-456, December 2015. Doi: 10.12720/jomb.4.6.451-456