Peer-Reviewed Articles
793 Methodology For Hyperspectral Band Selection
Peter Bajcsy and Peter Groves
Abstract Download
Full Article (members only)
While hyperspectral data are very rich in information, processing the hyperspectral
data poses several challenges regarding computational requirements, information
redundancy removal, relevant information identification, and modeling accuracy.
In this paper we present a new methodology for combining unsupervised and
supervised methods under classification accuracy and computational requirement
constraints that is designed to perform hyperspectral band (wavelength range)
selection and statistical modeling method selection. The band and method
selections are utilized for prediction of continuous ground variables using
airborne hyperspectral measurements. The novelty of the proposed work is
in combining strengths of unsupervised and supervised band selection methods
to build a computationally efficient and accurate band selection system.
The unsupervised methods are used to rank hyperspectral bands while the accuracy
of the predictions of supervised methods are used to score those rankings.
We conducted experiments with seven unsupervised and three supervised methods.
The list of unsupervised methods includes information entropy, first and
second spectral derivative, spatial contrast, spectral ratio, correlation,
and principal component analysis ranking combined with regression, regression
tree, and instance-based supervised methods. These methods were applied to
a data set that relates ground measurements of soil electrical conductiveity
with airborne hyperspectral image values. The outcomes of our analysis led
to a conclusion that the optimum number of bands in this domain is the top
four to eight bands obtained by the entropy unsupervised method followed
by the regression tree supervised method evaluation. Although the proposed
band selection approach is demonstrated with a data set from the precision
agriculture domain, it applies in other hyperspectral application domains.
803 Wavelets for Urban Spatial Feature
Discrimination: Comparisons with Fractal, Spatial Autocorrelation,
and Spatial Co-occurrence Approaches
Soe Win Myint, Nina Lam, and John M. Tyler
Abstract Download
Full Article (members only)
Traditional image processing techniques have proven inadequate for urban mapping
using high spatial resolution remote-sensing images. This study examined
and evaluated wavelet transforms for urban texture analysis and image classification
using high spatial resolution ATLAS imagery. For the purpose of comparison
and to evaluate the effectiveness of the wavelet approaches, two different
fractal approaches (isarithm and triangular prism), spatial autocorrelation
(Moran's I and Geary's C), and spatial co-occurrence matrix of the selected
urban classes were examined using 65 x 65, 33 x 33, and 17 x 17 samples with
a pixel size of 2.5 m. Results from this study suggest that a multi-band
and multi-level wavelet approach can be used to drastically increase the
classification accuracy. The fractal techniques did not provide satisfactory
classification accuracy. Spatial autocorrelation and spatial co-occurrence
techniques were found to be relatively effective when compared to the fractal
approaches. It can be concluded that the wavelet transform approach is the
most accurate of all four approaches.
813 A Comparison of AVIRIS and Landsat
for Land Use Classification at the Urban Fringe
Rutherford V. Platt and Alexander F.H. Goetz
Abstract Download
Full Article (members only)
In this study we tested whether AVIRIS data allowed for improved land use classification
over synthetic Landsat ETM? data for a location on the urban-rural fringe
of Colorado. After processing the AVIRIS image and creating a synthetic Landsat
image, we used standard classification and post-classification procedures
to compare the data sources for land use mapping. We found that, for this
location, AVIRIS holds modest, but real, advantages over Landsat for the
classification of heterogeneous and vegetated land uses. Furthermore, this
advantage comes almost entirely from the large number of sensor spectral
bands rather than the high Signal-to-Noise Ratio (SNR).
821 A SPLIT Model for Extraction of Subpixel
Impervious Surface Information
Yeqiao Wang and Kevin Zhang
Abstract Download
Full Article (members only)
This paper introduces a Subpixel Proportional Land cover Information Transformation
(SPLIT) model to extract proportions of impervious surfaces in urban and
suburban areas. High spatial resolution airborne Digital Multispectral Videography
(DMSV) data provided subpixel information for Landsat TM data. The SPLIT
model employed a Modularized Artificial Neural Network (MANN) to integrate
multi-sensor remote sensing data and to extract proportions of impervious
surfaces and other types of land cover within TM pixels. Through a control
unit, the MANN was able to decompose a complex task into multiple subtasks
by using a group of sub-networks. The SPLIT model identified spectral relations
between TM pixel values and the corresponding DMSV subpixel patterns. The
established relationship allows extrapolation of the SPLIT model to the areas
beyond DMSV data coverage. We applied five intervals, i.e., <20 percent,
21 to 40 percent, 41 to 60 percent, 61 to 80 percent, and >81 percent,
to map the subpixel proportions of land cover types. We extrapolated the
SPLIT model from training sites that have both TM and DMSV coverage into
the entire DuPage County with TM data as the input. The extrapolation received
82.9 percent overall accuracy for the extracted proportions of urban impervious
surface.
829 Development of a 2001 National Landcover
Database for the United States
Collin Homer, Chengquan Huang, Limin Yang, Bruce Wylie, and Michael Coan
Abstract Download
Full Article (members only)
Multi-Resolution Land Characterization 2001 (MRLC 2001) is a second-generation
Federal consortium designed to create an updated pool of nation-wide Landsat
5 and 7 imagery and derive a second-generation National Land Cover Database
(NLCD 2001). The objectives of this multi-layer, multi-source database are
two fold: first, to provide consistent land cover for all 50 States, and
second, to provide a data framework which allows flexibility in developing
and applying each independent data component to a wide variety of other applications.
Components in the database include the following: (1) normalized imagery
for three time periods per path/row, (2) ancillary data, including a 30 m
Digital Elevation Model (DEM) derived into slope, aspect and slope position,
(3) per pixel estimates of percent imperviousness and percent tree canopy,
(4) 29 classes of land cover data derived from the imagery, ancillary data,
and derivatives, (5) classification rules, confidence estimates, and metadata
from the land cover classification. This database is now being developed
using a Mapping Zone approach, with 66 Zones in the continental United States
and 23 Zones in Alaska. Results from three initial mapping Zones show single-pixel
land cover accuracies ranging from 73 to 77 percent, imperviousness accuracies
ranging from 83 to 91 percent, tree canopy accuracies ranging from 78 to
93 percent, and an estimated 50 percent increase in mapping efficiency over
previous methods. The database has now entered the production phase and is
being created using extensive partnering in the Federal government with planned
completion by 2006.
841 Determination of Burnt Scars Using
Logistic Regression and Neural Network Techniques from a Single Post-Fire
Landsat-7 ETM+ Image
Ruiliang Pu and Peng Gong
Abstract Download
Full Article (members only)
Using logistic regression (LR) and artificial neural networ (NN) algorithms,
probabilities of presence/absence (p/a) of burned scars were calculated from
post-fire Landsat 7 Enhanced Thematic Mapper plus (ETM?) images in mountainous
areas of northern California. The discriminating power of six original TM
bands (TM bands 1 through 5 and band 7) and five vegetation indices between
burned and unburned areas were analyzed. The LR and NN techniques were applied
to two study sites with varied topography. We evaluated the performance of
both methods in predicting burned scars based on predictive accuracy, uncertainty
index, and computation time. The experimental results indicate that (1) the
LR is more efficient than the NN in predicting burned scars, but both techniques
can produce similar and acceptable prediction accuracy (overall average accuracy
greater than 97 percent for both methods at the two study sites) of p/a of
burned areas; (2) among all six original TM bands and five vegetation indices,
original TM4 and TM7 and NDVI1 (TM4, TM7) and NDVI2 (TM4, TM3) exhibit the
highest discrimination between burned scars and unburned vegetation areas;
and (3) the predictive accuracy produced with samples from the shaded and
shadowed areas is lower than that from the sunlit areas.
851 Mobile GIS and Speech Recognition
Andrew Hunter and Naser El-Sheimy
Abstract Download
Full Article (members only)
The research investigated whether a Mobile Geographic Information System (MGIS)
incorporating speech recognition was a viable tool for locating defects in
the streetscape. The Geography Markup Language for encoding spatial information
was used to implement an application schema for street condition surveys.
Speech accuracy exceeded 95% in environments that were quiet or constantly
loud. However, for tests where the noise level varied, recognition accuracy
plummeted to 58 per-cent. Accuracy of captured defects was determined while" standing," "walking," "cycling," and "driving." Errors
ranged from 0.27 m to 12.49 m at the 95 percent confidence interval. A web-based
questionnaire indicated that municipal geographic information users are unhappy
with the quality of their data, and as yet, do not require data in real-time.
Future research involves investigating alternative ways of capturing spoken
commands, the effect that mobile computing has on the cognitive abilities
of the user, and wireless connectivity required for real time access to spatial
data.
| Top | Home |