METHODS
Information in this study was derived from the 2002 aerial orthophotography (scale 1:4000), THAICHOTE
panchromatic imagery (2m.resolution) acquired in December 2011 and national forest reserve area (1:50000)
available from Royal forest department. The Thaichote images were geo-metrically corrected using the aerial
orthophotography and the nearest neighbor resampling method.
On screen digitizing method of the aerial orthophotography and the Thaichote images was used to identify the
LCLU and rubber tree plantation in the study area. Directly observable features of the high resolution images
provide identification the LCLU and rubber tree growth with different ages rubber tree. The mature and middle
ages rubber tree could be discriminated by the difference in plant canopy and pattern. The young rubber tree
plantation could be resolved through a pattern of planting space of 4*6-5*7m. The young rubber tree plantation is
mostly composed of the mixed pixels of bare ground, grass, weeds, small shrub and the young rubber trees. The
ensuing mixtures pose a difficulty in resolving the young rubber tree when using digital classification method.
Over 98 training sites of different classes and tree ages were investigated to establish the ground truth of the covers
and site characteristics. To ensure the observation, estimate and measurement of the tree stands was carried out. The
obtained information was checked against the classified images. Kappa coefficient was calculated to validate the
resulted maps.
To address the expansion areas of rubber tree in conservation forest (CF) and paddy land. The 2002 LCLU layer
was overlaid with the 2011 LCLU layer, yielding the rubber tree encroachment on paddy land. The encroachment
of the plantation on conservation forest was performed by overlay the2011 LCLU layer with the conservation area.