Peer-Reviewed Articles
789 An Accuracy Index with Positional and Thematic
Fuzzy Bounds for Land-use / Land-cover Maps
Stéphane Couturier, Jean-François Mas, Gabriela Cuevas,
Jorge Benítez, Álvaro Vega-Guzmán, and Valdemar Coria-Tapia
Abstract Download
Full Article
This paper proposes a comprehensive framework for the
accuracy assessment of taxonomically diverse LULC maps.
A widely accepted difficulty in assessing such maps is
associated with the vagueness in the interpretation of
complex landscapes. For every class of the map, this
method quantified the thematic and positional fuzziness of
accuracy, induced by this difficulty. The labeling protocol
consisted of a fuzzy comparison between the map and a
reference maplet, for which degrees of positional and
thematic tolerance can be user-defined. The construction
of reference maplets permitted a flexible analysis (comparable
with the assessment of other maps) of the positional
fuzziness of the reference dataset and of the vagueness of
the assessment process, while the alternate evaluation
protocol, based on traditional point like data collection,
did not allow such analysis.
807 GIS Analysis of Global Impacts from Sea Level Rise
Xingong Li, Rex J. Rowley, John C. Kostelnick, David
Braaten, Joshua Meisel, and Kalonie Hulbutta
Abstract Download
Full Article
Future sea level rise caused by climate change would disrupt
the physical processes, economic activities, and social systems
in coastal regions. Based on a hypothetical global sea level
increase of one to six meters, we developed GIS methods to
assess and visualize the global impacts of potential inundation
using the best available global datasets. After susceptible
areas were delineated, we estimated that the size of the areas
is between 1.055 (one meter) to 2.193 million km2 (six meters).
Population in the susceptible areas was estimated to range
from 108 (one meter) to 431 million (six meters) people.
Among the seven land-cover types in the susceptible areas,
forest and grassland account for more than 60 percent for all
the increments of sea level rise. A suite of interactive visualization
products was also developed to understand and
communicate the ramifications of potential sea level rise.
819 Forest Type Mapping using Object-specific Texture
Measures from Multispectral Ikonos Imagery: Segmentation
Quality and Image Classification Issues
Minho Kim, Marguerite Madden, and Timothy A. Warner
Abstract Download
Full Article
This study investigated the use of a geographic object-based
image analysis (GEOBIA) approach with the incorporation
of object-specific grey-level co-occurrence matrix (GLCM)
texture measures from a multispectral Ikonos image for
delineation of deciduous, evergreen, and mixed forest
types in Guilford Courthouse National Military Park, North
Carolina. A series of automated segmentations was produced
at a range of scales, each resulting in an associated
range of number and size of objects (or segments). Prior to
classification, the spatial autocorrelation of each segmentation
was evaluated by calculating Moran’s I using the
average image digital numbers (DNs) per segment. An initial
assumption was made that the optimal segmentation scales
would have the lowest spatial autocorrelation, and conversely,
that over- and under-segmentation would result
in higher autocorrelation between segments. At these
optimal segmentation scales, the automated segmentation
was found to yield information comparable to manually
interpreted stand-level forest maps in terms of the size and
number of segments. A series of object-based classifications
was carried out on the image at the entire range of segmentation
scales. The results demonstrated that the scale of
segmentation directly influenced the object-based forest
type classification results. The accuracies were higher for
classification of images identified from a spatial autocorrelation
analysis to have an optimal segmentation, compared
to those determined to have over- and under-segmentation.
An overall accuracy of 79 percent with a Kappa of 0.65
was obtained at the optimal segmentation scale of 19.
The addition of object-specific GLCM multiple texture
analysis improved classification accuracies up to a value
of 83 percent overall accuracy and a Kappa of 0.71 by
reducing the confusion between evergreen and mixed forest
types. Although some misclassification still remained
because of local segmentation quality, a visual assessment
of the texture-enhanced GEOBIA classification generally
agreeable with manually interpreted forest types.
831 Developing Collaborative Classifiers using an
Expert-based Model
Giorgos Mountrakis, Raymond Watts, Lori Luo, and Jida
Wang
Abstract Download
Full Article
This paper presents a hierarchical, multi-stage adaptive
strategy for image classification. We iteratively apply
various classification methods (e.g., decision trees, neural
networks), identify regions of parametric and geographic
space where accuracy is low, and in these regions, test
and apply alternate methods repeating the process until
the entire image is classified. Currently, classifiers are
evaluated through human input using an expert-based
system; therefore, this paper acts as the proof of concept
for collaborative classifiers. Because we decompose the
problem into smaller, more manageable sub-tasks, our
classification exhibits increased flexibility compared to
existing methods since classification methods are tailored to
the idiosyncrasies of specific regions. A major benefit of our
approach is its scalability and collaborative support since
selected low-accuracy classifiers can be easily replaced with
others without affecting classification accuracy in high
accuracy areas. At each stage, we develop spatially explicit
accuracy metrics that provide straightforward assessment
of results by non-experts and point to areas that need
algorithmic improvement or ancillary data. Our approach
is demonstrated in the task of detecting impervious surface
areas, an important indicator for human-induced alterations
to the environment, using a 2001 Landsat scene from
Las Vegas, Nevada.
Color Figure 1 (Adobe PDF Format 2.65Mb)
Color Figure 7 (Adobe PDF Format 11.3Mb)
845 Automated Urban Delineation from Landsat
Imagery Based on Spatial Information Processing
Bert Guindon and Ying Zhang
Abstract Download
Full Article
Methodologies are developed and assessed for the
delineation of urban areas on moderate resolution satellite
images based on line feature extraction and processing. It
is argued that a simple two component urban/rural road
network model and a preliminary estimate of the extent of
urban cover in a satellite scene can be used to quantify
processing settings, such as line thresholds. As a result,
urban delineation can be fully automated. Tests have been
conducted with Landsat Thematic Mapper data of a candidate
set of cities that exemplify the spectrum of North
American urban landscapes. Producer accuracies in the
range 60 to 80 percent are consistently for a variety of urban
landscapes. While these performance levels are lower than
those previously reported for small area, city-specific
studies, they are comparable to those previously achieved in
large-area, approaches (e.g., the National Land Cover
Dataset initiative) that rely primarily on spectral attributes.
It is concluded that better integrated use of spectral-spatial
data processing have the potential to lead to improved
operational urban mapping accuracies.
859 Traffic Monitoring using Very High Resolution
Satellite Imagery
Siri Øyen Larsen, Hans Koren, and Rune Solberg
Abstract Download
Full Article
Very high resolution satellite images allow automated
monitoring of road traffic conditions. Satellite surveillance
has several obvious advantages over current methods,
which consist of expensive single-point measurements made
from pressure sensors, video surveillance, etc., in/or close to
the road. The main limitation of using satellite surveillance
is the time resolution; the continuously changing traffic
situation must be deduced from a snapshot image. In
cooperation with the Norwegian Road Authorities, we have
developed an approach for detection of vehicles in Quick-Bird images. The algorithm consists of a segmentation step
followed by object-based maximum likelihood classification.
Additionally, we propose a new approach for prediction
of vehicle shadows. The shadow information is used as a
contextual feature in order to improve classification. The
correct classification rate was 89 percent, excluding noise
samples. The proposed method tends to underestimate the
number of vehicles when compared to manual counts and
in-road equipment counts.
871 Object-based Detection and Classification of
Vehicles from High-resolution Aerial Photography
Ashley C. Holt, Edmund Y.W. Seto, Tom Rivard, and Peng
Gong
Abstract Download
Full Article
Vehicle counts and truck percentages are important input
variables in both noise pollution and air quality models,
but the acquisition of these variables through fixed-point
methods can be expensive, labor-intensive, and provide
incomplete spatial sampling. The increasing availability and
decreasing cost of high spatial resolution imagery provides
an opportunity to improve the descriptive ability of traffic
volume analysis. This study describes an object-based classification
technique to extract vehicle volumes and vehicle
type distributions from aerial photos sampled throughout a
large metropolitan area. We developed rules for optimizing
segmentation parameters, and used feature space optimization
to choose classification attributes and develop fuzzy-set
memberships for classification. Vehicles were extracted from
street areas with 91.8 percent accuracy. Furthermore, separation
of vehicles into classes based on car, medium-sized
truck, and buses/heavy truck definitions was achieved with
87.5 percent accuracy. We discuss implications of these
results for traffic volume analysis and parameterization
of existing noise and air pollution models, and suggest
future work for traffic assessment using high-resolution
remotely-sensed imagery.
Color Figure 6 (Adobe PDF Format 13.9Mb)