forest, with many and larger gaps between trees, which led to
confusion of dead trees with such cover types as bare ground,
senescent grass, and decadent shrubs. The forests of Palomar
and Volcan consist of closed canopy groves, so the dead tree to
background (live tree) contrast is more distinct. These general
trends are also apparent for the spatial contextual classifier
results, although the accuracy differences are only statistically
significant between the Volcan and Laguna sites. Other factors
that may have influenced differences between study sites could
be differences in tree types, the physiological manifestation of
mortality, and the nature of the terrain within the study sites.
Summary and Conclusions
Spatial contextual machine learning (
ANN
) and object-based
approaches to image-based mapping of dead tree objects
yielded useful maps of dead tree locations, with the majority
of dead tree objects correctly mapped and with reasonable
commission errors. Procedures utilized to optimize each prod-
uct, such as masking techniques and the addition of spectral
transforms improved the classification products by 5 percent
to 10 percent, with an 8 percent to 13 percent reduction in
commission error. When comparing the two approaches, dead
conifer tree class accuracies are higher overall for the spatial
contextual approach than the object-based approach. The over-
all accuracies for both methods ranged between 30 percent to
90 percent for the dead tree class, but overall, the accuracies
were significantly higher for the spatial-contextual approach.
From a qualitative perspective, the spatial-contextual
approach produced greater object quality, and was a much
easier program to learn and implement. Manual editing of
map products helped to bring the classifications up to more
acceptable levels of accuracy for further analysis.
Accurate and timely tree mortality maps are needed for forest
management, especially in areas prone to drought and frequent
fires such as the montane areas of San Diego County. Digital im-
age classification is a viable method for producing these maps
from freely available high-spatial resolution aerial orthoimagery.
However, it is important to select appropriate image classifica-
tion techniques and spectral inputs to obtain the highest map
accuracy possible. Although tree mortality is inherently difficult
to map, two approaches have been demonstrated to yield maps
of dead tree objects such that the majority of dead tree objects
are detected and with reasonable commission error. Even
though some degree of visual interpretation and editing is likely
required to yield final products of sufficient reliability for exam-
ining spatial-temporal patterns of tree mortality, these semi-
automated image classification procedures are beneficial, in that
they are much more efficient at mapping dead tree objects than
using manual interpretation and digitizing methods only.
References
Al-Khudhairy, D.H.A., I. Caravaggi, and S. Giada, 2005. Structural
damage assessments from Ikonos data using change detection,
object-oriented segmentation, and classification techniques,
Photogrammetric Engineering & Remote Sensing
, 71(7):825–837.
doi:10.14358/PERS.71.7.825.
Anderson, J.R., E.E. Hardy, J.T. Roach, and R.E. Witmer, 1976. A land
use and land cover classification system for use with remote
sensor data, US Government Printing Office, 964:27.
Atkinson, P.M., and A.R.L. Tatnall, 1997. Introduction to neural
networks in remote sensing,
International Journal of Remote
Sensing
, 18(4):699–709. doi:10.1080/014311697218700.
Benz, U.C., P. Hofmann, G. Willhauck, I. Lingenfelder, and M.
Heynen, 2004. Multi-resolution, object-oriented fuzzy analysis
of remote sensing data for GIS-ready information,
ISPRS Journal
of Photogrammetry & Remote Sensing
, 58(3–4):239–258.
doi:10.1016/j.isprsjprs.2003.10.002.
Blaschke, T., 2010. Object based image analysis for remote sensing,
ISPRS Journal of Photogrammetry and Remote Sensing
, 65(1):2–
16. doi:10.1016/j.isprsjprs.2009.06.004.
Blaschke, T., and J. Strobl, 2001. What’s wrong with pixels? Some
recent developments interfacing remote sensing and GIS,
GeoBIT/GIS
, 6(01):12–17.
Chubey, M.S., S.E. Franklin, and M.A. Wulder, 2006. Object-based
analysis of Ikonos-2 imagery for extraction of forest inventory
parameters,
Photogrammetric Engineering & Remote Sensing
,
72(4):383–394. doi:10.14358/PERS.72.4.383.
Erikson, M., 2003. Segmentation of individual tree crowns in colour
aerial photographs using region growing supported by fuzzy
rules,
Canadian Journal of Forest Research
, 33(8):1557–1563.
doi:10.1139/x03-062.
Flanders, D., M. Hall-Beyer, and J. Pereverzoff, 2003. Preliminary
evaluation of eCognition object-based software for cut block
delineation and feature extraction,
Canadian Journal of Remote
Sensing
, 29(4):441–452. doi:10.5589/m03-006.
Foody, G.M., and M.K. Arora, 1997. An evaluation of some factors
affecting the accuracy of classification by an artificial neural
network,
International Journal of Remote Sensing
, 18(4):799–
810. doi:10.1080/014311697218764.
Gitelson, A.A., Y.J. Kaufman, R. Stark, R., and D. Rundquist, 2002.
Novel algorithms for remote estimation of vegetation fraction,
Remote Sensing of Environment
, 80(1):76–87. doi:10.1016/
S0034-4257(01)00289-9.
Graetz, R.D., 1990. Remote sensing of terrestrial ecosystem structure:
An ecologist’s pragmatic view,
Remote Sensing of Biosphere
Functioning
(R.J. Hobbs and H.A. Mooney, editors) Springer-
Verlag, New York, 312 p.
Guo, Q., M. Kelly, P. Gong, and D. Liu, 2007. An object-based
classification approach in mapping tree mortality using high
spatial resolution imagery,
GIScience and Remote Sensing
,
44(1):24–47. doi:10.2747/1548-1603.44.1.24.
Gustafson, E.J., 1998. Minireview: Quantifying Landscape Spatial
Pattern: What Is the State of the Art?,
Ecosystems
, 1(2):143–156.
doi:10.1007/s100219900011.
Huete, A.R., 1989. Soil influences in remotely sensed vegetation-
canopy spectra,
Theory and Applications of Optical Remote
Sensing
(G. Asrar, editor), John Wiley and Sons, Inc., pp. 107-141.
Im, J., J.R. Jensen, and J.A. Tullis, 2008. Object-based change detection
using correlation image analysis and image segmentation,
International Journal of Remote Sensing
, 29(2):399–423.
Jensen, J., 2006.
Remote Sensing of the Environment: An Earth
Resource Perspective
, Second edition, Prentice Hall Series of
Geographic Information Science., Prentice Hall, New Jersey, 544 p.
Jordan, C.F., 1969. Derivation of leaf-area index from quality of light
on the forest floor,
Ecology
, 50(4):663–666. doi:10.2307/1936256.
Kaiser, J.V., D.A. Stow, and L. Cao, 2004. Evaluation of remote sensing
techniques for mapping transborder trails,
Photogrammetric
Engineering & Remote Sensing
, 70(12):1441–1447.
Kerr, J.T., and M. Ostrovsky, 2003. From space to species: Ecological
applications for remote sensing,
Trends in Ecology and
Evolution
, 18(6):299–305. doi:10.1016/S0169-5347(03)00071-5.
Lang, S., E. Schöpfer, and T. Langanke, 2009. Combined object-
based classification and manual interpretation - Synergies for
a quantitative assessment of parcels and biotopes,
Geocarto
International
,
24
(2):99–114. doi:10.1080/10106040802121093.
Lippitt, C.D., L.L. Coulter, M. Freeman, J. LaMantia-Bishop, W. Pang,
and D.A. Stow, 2012. The effect of input data transformations on
object-based image analysis,
Remote Sensing Letters
, 3(1):21–29.
doi:10.1080/01431161.2010.531059.
Mallinis, G., N. Koutsias, M. Tsakiri-Strati, and M., Karteris, 2008.
Object-based classification using QuickBird imagery for
delineating forest vegetation polygons in a Mediterranean test
site,
ISPRS Journal of Photogrammetry and Remote Sensing
,
63(2):237–250. doi:10.1016/j.isprsjprs.2007.08.007.
Poznanovic, A.J., M.J. Falkowski, A.L. Maclean, A.M.S. Smith, and
J.S. Evans, 2014. An accuracy assessment of tree detection
algorithms in juniper woodlands,
Photogrammetric Engineering
& Remote Sensing
, 80(7):627–637. doi:10.14358/PERS.80.7.627.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
July 2016
579
