American Journal of Applied Sciences 7 (4): 486-492, 2010
ISSN 1546-9239
© 2010Science Publications
486
A New Approach in Individual Tree Counting of Nyatoh (Palaquium spp.)
Using an Airborne Spectroradiometer
Kamaruzaman Jusoff
Tropical Forest Airborne Observatory (TropAIR), Faculty of Forestry, Universiti Putra Malaysia,
43400 Serdang, Selangor. Malaysia
Abstract: Problem statement: In Malaysia inadequate data are available to make decisions regarding
the sustainable management of Nyatoh (Palaquium spp.). Factual data and estimation of Nyatoh in the
lowland dipterocarp forest resources are still lacking or, if available, are not complete, reliable or up to
date. The collection and compilation of information and updating of existing data in near real time is a
priority. Nyatoh is one of the important and popular timber species for furniture, solid doors and
planks. Approach: In order to determine Nyatoh’s latest species count in the Federal Territory (FT)’s
Permanent Reserved Forest, a study on the distribution of Nyatoh in Bukit Nanas Forest Reserve (FR),
FT was conducted using an airborne hyperspectral remote sensing technique with an objective to
identify, count and map the species of Nyatoh in Bukit Nanas FR, FT. A False Color Composite (FCC)
airborne hyperspectral image of the study area was used with a Sobel 3×3 filter and Spectral Angle
Mapper (SAM) to classify and separated the individual Nyatoh from other tree species within Bukit
Nanas FR. Results: A thematic map of individual Nyatoh spatial distribution in Bukit Nanas FR was
developed showing a total of 36 individual Nyatoh trees in a 2.7 h study plot. There were 34 trees
confirmed available on the ground with a mapping accuracy of 94.4%. Conclusion: The AISA
airborne spectroradiometer data is capable of mapping and counting individual tropical timber trees
especially Palaquium spp. with an acceptable accuracy of more than 90%. However, in order to
improve the mapping accuracy, it is recommended that a higher spatial resolution image data is
necessary in addition to a higher precision field GPS usage. This approach may be well integrated in
the GIS-based decision support system to better manage and conserve such depleting timber species in
the FT’s state land and permanent reserved forests.
Key words: Nyatoh (Palaquium spp.), species count, tree mapping, airborne spectroradiometer,
spectral signature
INTRODUCTION
Forest mapping and timber inventory of the mixed
dipterocarp forests is one of the major applications for
Malaysia’s new airborne hyperspectral sensor. Airborne
hyperspectral remote sensing can provide forest
information products for applications in forest
inventory, forest chemistry and for some Kyoto
Protocol information products. Through several
projects, it has been demonstrated that airborne
hyperspectral sensing can provide accurate maps of hill
and mountain mixed dipterocarp individual forest
species (Jusoff, 2006; 2007; 2008a; 2008b). High
correlations have been demonstrated between field
measurements of the image versus field spectra estimates
derived from hyperspectral sensing (Jusoff, 2008c;
2009a, 2009b, 2009c; 2009d; Jusoff et al., 2009;
Mohd Hasmadi et al., 2010; Jusoff et al., 2010). Data for
these experiments have included the UPM’s-AISA data
for different applications (Jusoff and Pathan, 2009;
Jusoff, 2009a; 2009b; 2009c; 2009d; 2009e; Jusoff and
Yusoff, 2009). To achieve operational accuracies of
90% or better for individual tropical forest species
classification, it is essential that the airborne
hyperspectral data are well calibrated and tested over
several test sites in Malaysia. This study will report on
the latest research state of airborne hyperspectral
sensing of a mixed dipterocarp lowland forests in Bukit
Nanas forest reserve, federal territory, Kuala Lumpur
and present the airborne hyperspectral sensor
approaches needed for such timber inventory at the
species level applications.
Forest inventory information is obtained by
measuring and assessing the trees and various
characteristics of the forest. The information may be
obtained from measurements taken on the ground or on
remotely sensed imagery. When the measurements are
taken for the entire forest, the inventory is a complete