Peer-Reviewed Articles
973 MODIS-based Change Detection for Grizzly Bear
Habitat Mapping in Alberta
Alysha D. Pape and Steven E. Franklin
Abstract Download
Full Article
Coarse resolution data from the Moderate Resolution
Imaging Spectroradiometer (MODIS) was used to test the
effectiveness of 250 m data to detect forest disturbances and
update an existing, large-area (150,000 km2), 30 m Landsat
ETM+ and TM land-cover map product used in Grizzly Bear
(Ursus arctos) habitat analysis. A Landsat-derived polygon
layer was applied to the MOD13Q1 data product to create a
polygon-based, mean NDVI time series (2000 to 2005). Image
differencing of the dataset produced multiple-scale layers of
change including a two-date, five-year change and a fiveyear
composite of annual changes. Accuracy assessments
based on available GIS data showed an overall accuracy as
high as 59 percent. Results also show that disturbance
patches larger than 15 ha were represented with an accuracy
of 75 percent or higher. This offers an alternative to
higher spatial resolution data for the identification of larger
features and also provides general change information for
those areas that may be suitable for analysis with higher
spatial resolution data.
987 Quantitative Mapping of Hydrodynamic Vegetation
Density of Floodplain Forests Under Leaf-off Conditions
Using Airborne Laser Scanning
Menno W. Straatsma
Abstract Download Full Article
In this paper a method is presented to extract hydrodynamic
vegetation density from airborne laser scanner data, relevant
for exceedance levels of embankments of lowland areas. Two
indices to predict vegetation density from the laser data were
considered: (a) Percentage Index (PI) of points in the height
interval inundated by the water, and (b) the Vegetation Area
Index (VAI) that corrects for occlusion from the crown area.
A computer simulation, using a digital forest model, showed
a sensitivity of the indices for laser pulses that were sent out,
but not detected by the laser receiver. The locations of these
invalid points were therefore reconstructed. Two different
assumptions were tested to assign new coordinates to these
so-called invalid points. Percentage Index, with the invalid
points reconstructed by means of thresholding the point
density ratio, proved the best predictor (R2 = 0.66) of vegetation
density of deciduous floodplain forests under winter
conditions..
999 Urban Change Detection Based on Coherence and
Intensity Characteristics of SAR Imagery
Mingsheng Liao, Liming Jiang, Hui Lin, Bo Huang,
and Jianya Gong
Abstract Download
Full Article
In this paper, an unsupervised change-detection approach
was proposed to detect new urban areas from multi-temporal
SAR images. The novelty of the proposed approach is the
joint use of coherence and intensity characteristics of SAR
imagery. The approach involves two main steps: (a) the
extraction of difference feature containing information
on changed areas, and (b) the unsupervised two-dimensional
(2D) thresholding. First, two difference features based on
the concepts of long-term coherence and backscattering
temporal variability are extracted from a series of multitemporal
SAR images. Then, the resulting features that
represent the INSAR signal temporal variability of changed
areas are merged, and a 2D thresholding technique based on
the maximum 2D Renyi’s entropy criterion is developed to
obtain the change-detection results. The effectiveness of the
proposed approach is confirmed with experimental results
obtained from a set of six ERS-1/2 SLC SAR images acquired
in Shanghai, China.
1007 Factors Affecting Spatial Variation of Classification
Uncertainty in an Image Object-based Vegetation
Mapping
Qian Yu, Peng Gong, Yong Q. Tian, Ruiliang Pu and Jun Yang
Abstract Download
Full Article
Much effort has been spent on examining the spatial
variation of classification accuracy and associated
factors on a per-pixel basis. In the past few years,
object-based classification has attracted growing interest.
This paper examines factors affecting the spatial variation
of classification uncertainty in an object-based
vegetation mapping. We studied six categories of factors
in an object-based classification: general membership,
topography, sample object density, spatial composition,
sample object reliability, and object features. First,
classification uncertainty (classification accuracy on a
per-case basis) is derived with a bootstrap method. Then,
six categories of factors are quantified by categorical or
continuous variables. In this step, the appropriate
radius for calculating the spatial composition metrics
of sample objects is also discussed. Finally, classification
uncertainty is modeled as a function of those factors
using a mixed linear model. The significant factors
are identified and their parameters are estimated from
restricted maximum likelihood fit. The modeling results
show that elevation, sample object size, sample object
reliability, sample object density, and sample spatial
composition significantly influence the object-based
classification uncertainty. Many of these factors
are closely related to the object-based approach. The
result of this study helps in understanding classification
errors and suggests further improvement of the
classification.
Color Figures (Adobe PDF format):
1019 Evaluating Neural Networks and Evidence Pooling
for Land Cover Mapping
M.J. Aitkenhead, S. Flaherty, and M.E.J. Cutler
Abstract Download
Full Article
The diversity of data sources, analysis methodologies, and
classification systems has led to a number of new techniques
for monitoring land-cover change. However, this wide choice
means that it is difficult to know which solution to choose.
A system capable of integrating the results of different
analyses and applying them to land-cover mapping would
therefore be extremely useful. This study investigates the use
of evidence pooling and neural networks in land-cover
mapping. Neural networks were used to classify land-cover
using evidence from spectral (Landsat-7 ETM1), textural, and
topographic information. Mapping was performed using
combinations of evidence source and evidence pooling
techniques. The best performance was achieved using all
available information with a method that summed evidence
directly instead of categorizing it. While the methodology
failed to reach the level of accuracy recommended elsewhere,
a comparison of the number of classes used with other
methods showed that the system performed better than these
approaches.
1033 Lidar-based Mapping of Forest Volume and
Biomass by Taxonomic Group Using Structurally
Homogenous Segments
Jan A.N. van Aardt, Randolph H. Wynne, and John A. Scrivani
Abstract Download
Full Article
This study evaluated the potential of an object-oriented
approach to forest type classification as well as volume and
biomass estimation using small-footprint, multiple return
lidar data. The approach was applied to coniferous, deciduous,
and mixed forest stands in the Virginia Piedmont,
U.S.A. A multiresolution, hierarchical segmentation
algorithm was applied to a canopy height model (CHM) to
delineate objects ranging from 0.035 to 5.632 ha/average
object. Per-object lidar point (per return height and intensity)
and CHM distributional parameters were used as input to a
discriminant classification of 2-class (deciduous-coniferous)
and 3-class (deciduous-coniferous-mixed) forest definitions.
Lidar point-height-based and CHM classifications yielded
overall accuracies of 89 percent and 79 percent, respectively.
Volume and biomass estimates exhibited differences of no
more than 5.5 percent compared to field estimates, while
showing distinctly improved precisions (up to 45.5 percent).
There were no significant differences between accuracies for
varying object sizes, which implies that reducing the lidar
point coverage would not affect classification accuracy.
These results lead to the conclusion that a lidar-based
approach to forest type classification and volume/biomass
assessment has the potential to serve as a single-source
inventory tool.
1045 Comparison of Spectral Analysis Techniques for Impervious
Surface Estimation Using Landsat Imagery
Fei Yuan, Changshan Wu, and Marvin E. Bauer
Abstract Download
Full Article
Various methodologies have been used to estimate and map
percent impervious surface area (%ISA) using moderate
resolution remote sensing imagery (e.g., Landsat Thematic
Mapper). There is, however, a lack of comparative analyses
among these methods. This study compares three major
spectral analysis techniques (regression modeling, regression
tree, and normalized spectral mixture analysis (NSMA)) for
continuous %ISA estimation using Landsat imagery for
1986 and 2002 for the seven-county Twin Cities Metropolitan
Area of Minnesota. Our study showed that all three
techniques demonstrate the capability for estimating %ISA
accurately, with RMSE ranging from 7.3 percent to 11 percent
and R2 of 0.90 to 0.96 for both years. Comparatively,
regression modeling and regression tree methods produced
similar results; however, both of them are highly dependent
on accurate masks to differentiate urban impervious
surfaces from bare soil. Within the urban mask, the
regression tree-based estimates were the most accurate. In
terms of time and cost, the NSMA approach is most efficient,
but it tends to underestimate the percent imperviousness for
highly developed areas. Findings from the study provide
guidance for the selection of %ISA estimation techniques
using moderate resolution remote sensing data, along with
information for further methodological improvements.