bul 2004
ne. fact
fferent
teristic
e. kind
ephala
y, the
on to
till” be
second
The
ly this
olution
results
es that
znizing
stics in
efect ‘is
broad
hysical
> target
lysis of
ach to
inalysis
were
ults in
roduce
lution
Lin Fig.
ucaena
ce in
perhaps
pear to
cause a
ifferent
spectral
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B7. Istanbul 2004
appearance, they all have very similar, if not exactly
the same, and unique spatial patterns (texture) and
are very distinct from other vegetation covers, such
as the farmlands and grasslands also shown in the
same image. As a result, after the refinement by
texture analysis on high resolution images, the
accuracy of Leucaena Leucocephala detection was
increased by at least 15%.
3.3 Future Improvement
First of all, as mentioned above, the spatial (texture)
analysis of high resolution images in this study was
accomplished through interactive analysis by
experienced human interpreters. The top priority
will be to computerize the process in order to
achieve more degree of automation. Several texture
analysis algorithms (e.g, GLCM and CDTM) have
been evaluated. In addition, because it requires a
significant amount of high resolution data to cover
the entire study area, hence the data volume may
become too large for a timely and efficient full-scale
texture analysis. Therefore, a texture analysis
algorithm and procedure with level of detail (LOD)
consideration is under development and will be
implemented to address this issue.
Secondly, although spectral analysis operating on
MNF transformed hyperspectral data can produce
reasonable results, there is still room for
improvement. Other spectral analysis techniques
shall be investigated to explore the possibility of
creating a more effective method for spectral
analysis. Also, the integration between spectral and
texture analysis phases should also be addressed in
order to streamline the overall procedure of the
system.
4 CONCLUSION
The invasion of alien plant species has caused
significant impact to local ecosystems and
biodiversity in Taiwan. To better understand the
situation and develop strategies to battle against the
deterioration of this problem, it is necessary to have
an accurate knowledge about the distribution and
spreading status and trend of the invasive plants.
This study demonstrates that the coupling of
spectral analysis of hyperspectral images and
texture analysis on high resolution satellite data is
an effective and economic approach to detect
specific plant in a mesoscale to large area. The
systematic method developed in this research first
applies spectral analysis to MNF transformed
hyperspectal satellite images according to selected
spectral features. The preliminary results are then
further improved with texture analysis on high
resolution images. Example from this study shows
that the combination of the two analysis phases
(spectral and texture) can produce a reasonably well
accuracy in discriminating Leucaena Leucocephala
from local vegetation covers in the Kenting National
Park and surrounding areas.
ACKNOWLEDGEMENTS
The authors would like to thank the National Science
Council of Taiwan for their partial support to this
research and the travel to the conference. (Project
No. NSC-92-2211- E-008- 051).
REFERENCES
Chen, C.-F., P.-F. Lee, F. Tsai, L.-H. Liang, S.- M. Hsu,
S.-L. Cheng, 2003, "Applications of Airborne Multi-
Spectral Scanner Images” (in Chinese), COA
Technology Research Report No. 922029, Council of
Agriculture, Taipei, Taiwan.
Cochrane, M. A., 2000, "Using vegetation reflectance
variability for species level classification of
hyperspectral data", /nt'l J. Remote Sensing, 21(10),
pp. 2075- 2087.
Jiang, M.-Y. L.-M. Xiu, 2000, "The wildization,
imapcts, and management of nonnative plants in
Taiwan" (in Chinese), The Symp. Bio-diversity and
Preservation, Taipei, Taiwan, pp. 399-411.
Laba, : M, F. Tsai, D. Ogurcak, .S. Smith, M. E.
Richmond, 2003, "Use of Derivative Analysis for the
Spectral Discrimination of Invasive Wetland Plant
Species", submitted to PE & RS.
Lai, M.-J., 1995. Directories of Botanic Species in
Taiwan . (in Chinese) Di-Jing Pub., Taipei Taiwan.
Lins, K. F. and R. Kleckner, 1996, "Land use and land
cover mapping in the in the United States: An
overview and history of the convept" in GAP
Analysis: A Landscape approach to Biodiversity
Planning, eds. J. M. Scott, T. Tear and F. Davis, pp.
57-065.
Liu, F-Y., M.-A. Chen, 2002, "The Impacts. of Alien
Plants to Native Vegetation Ecosystems in Kenting
National Park — an example of Leucaena
Leucocephala" (in Chinese), Conservation Research
Report No. 112, Kenting National Park, Taiwan.
McCormick, C. M., 1999, "Mapping Exotic Vegetation
in the Everglades using large-scale aerial
photographs", PE&RS, 65(2), pp. 179- 184.
Schmidt, K. S, A. K. Skidmore, 2003, "Spectral
discrimination of vegetation types in a coastal
wetland", Remote Sensing of Environ., vol. 85, pp.
92- 108.
Stein, B. A., S. R. Flack, (eds.), 1996, America's Least
Wanted: Alien Species Invasions of U.S. Ecosystems,
The Nature Conservacy, Arlington, Virginia USA.
Tsai, F. and W. D. Philpot, 1998, "Derivative Analysis
of Hyperspectral Data", Remote Sensing of Environ.,
66(1), pp. 41- 51.
Tsai, F. and W. D. Philpot, 2002, "A Derivative- Aided
Image Analysis System for Land- Cover
Classification", IEEE Trans. on Geoscienc and Remote
Sensing , 40(2), pp. 416-425.
