Peer-Reviewed Articles
1005 Detection of Yellow Starthistle through Band Selection and Feature Extraction from Hyperspectral Imagery
Xin Miao, Peng Gong, Sarah Swope, Ruiliang Pu, Raymond Carruthers, and Gerald L. Anderson
Abstract Download
Full Article
To effectively display hyperspectral imagery for visualization
purposes, the three RGB channels should be selected or
extracted from a hyperspectral image under the criteria of
maximum information or maximum between-class separability.
Seven band selection (OIF, SI, CI, divergence, transformed
divergence, B-distance, JM-distance) and five feature extraction
(principal component analysis, linear discriminant
analysis, class-based PCA, segmented PCT (SPCT), independent
component analysis) methods and their variations are
examined and compared using CASI hyperspectral imagery
with the goal of detecting Centaurea solstitialis (yellow
starthistle or YST), an invasive weed, in an annual grassland
in California. Three indicators, information index (Infodex),
separability index (Sepadex) and average correlation
coefficient (ACC) are proposed to evaluate the quality of the
generated images. The results suggest that both the combination
of the three SPCT channels and the combination of the
second PCA channel with the positive and negative of the
first LDA channels (PCA2, LDA1, -LDA1) can enhance our
ability to visualize the distribution of YST in contrast to the
surrounding vegetation.
1017 Improving Pixel-based VHR Land-cover Classifications of Urban Areas with Post-classification Techniques
Tim Van de Voorde, William De Genst, and Frank Canters
Abstract Download
Full Article
In this paper, three post-classification techniques are
proposed to improve the information content, thematic
accuracy, and spatial structure of pixel-based classifications
of complex urban areas. A shadow-removal technique based
on a neural network that was trained using the output of a
soft classification is proposed to assign shadow pixels to
meaningful land-cover classes. Knowledge-based rules are
suggested to correct wrongly classified pixels and to improve
the overall accuracy of the land-cover map. Finally, a
region-based filter is applied to reduce high-frequency
structural clutter. The three techniques were successfully
applied to a pixel-based classification of a QuickBird image
covering the city of Ghent, Belgium, improving the kappa
index-of-agreement from 0.82 to 0.86 and transforming the
shadow pixels into meaningful land-cover information.
1029 Spectral Matching Techniques to Determine Historical Land-use/Land-cover and Irrigated Areas using Time-series 0.1 degree AVHRR Pathfinder Datasets
P.S. Thenkabail, P. GangadharaRao, T.W. Biggs, M. Krishna, and H. Turral
Abstract Download
Full Article
This study established spectral matching techniques (SMTs)
to determine land-use and land-cover (LULC) and irrigated
area classes from historical time-series (HTS-LULC) AVHRR
0.1-degree pathfinder satellite sensor data. The approach
for HTS-LULC mapping and characterization was to develop“target” spectra from: (a) Recent Time Series for which LULC
and irrigated area classes (RTS-LULC) were mapped using
extensive ground-truth data, and (b) ideal locations, which
are known endmembers even during historical time-periods
of interest, as determined based on existing knowledge base
including agricultural census data. The HTS-LULC for the
period of 1982 to 1985 and RTS-LULC for the period of 1996
to 1999 were established using monthly continuous timeseries
AVHRR mega-file data of 192 bands (48 months * 4
AVHRR bands per month) each for the HTS and RTS time
periods. The study was conducted in the Krishna river basin
(India), which has a large area (267,088 km2) with numerous
irrigation projects and high population density.
The quantitative and qualitative SMTs were used to identify and label HTS LULC classes. The identification and labeling process begins with qualitative spectral matching technique which visually matches the time-series NDVI spectra of known RTS-LULC classes and/or ideal endmember classes with time-series spectra of HTS-LULC classes. This helps identify classes of similar spectral characteristics in terms of shape and magnitude over time. The quantitative SMTs involved: (a) spectral correlation similarity (SCS), as a shape measure, (b) Euclidian distance (Ed), as distance measure, (c) spectral similarity value (SSV) as a combination of shape and distance measure, and (d) modified spectral angle similarity (MSAS) as a hyperangle measure. The quantitative and qualitative SMT methods and techniques lead to assigning HTS-LULC classes that match RTS-LULC names. In all, an aggregated seven HTS-LULC that were spectrally similar to the seven RTS-LULC classes and/or ideal endmember classes were identified and labeled. The SSV was the best method, followed by SCS.
1041 Adaptive Correlation Analysis With Non-Overlapping Imagery Indication
Frank Crosby
Abstract Download
Full Article
Many traditional image registration algorithms assume that
the shared structures in each of two images are a significant
portion of the image. However, in some remote sensing
applications, the overlap can be small or even nonexistent.
The registration method developed in this paper can assess
the existence of a common area and identify the alignment
between two images when the common area is a small
percentage of the images. Conventional correlation is used
as an input to an adaptive target detector designed for
registration. The output of the detector is then modeled with
a statistical distribution to find the significance of the
alignment point, which generally indicates if any overlap
was found. The method is applied to two data sets with
varying amounts of overlap. The data is low contrast, midwave
infrared, which is particularly challenging for traditional
registration algorithms and infrared imagery. The
results show that this method produces accurate registration
in a variety of cases where traditional correlation fails.
1049 Automatic Extraction of Main Road Centerlines from High Resolution Satellite Imagery Using Hierarchical Grouping
Xiangyou Hu and Vincent Tao
Abstract Download
Full Article
Automatic road centerline extraction from high-resolution
satellite imagery has gained considerable interest recently due
to the increasing availability of commercial high-resolution
satellite images. In this paper, a hierarchical grouping strategy
is proposed to automatically extract main road centerlines
from high-resolution satellite imagery. Here hierarchical
grouping means that, instead of grouping all segments at once,
the selective segments are grouped gradually, and multiple
clues are closely integrated into the procedure. By this means,
the computational cost can be reduced significantly. Through
the stepwise grouping, the detected fragmented line segments
eventually form the long main road lines. The proposed
method has been tested and validated using several Ikonos
and QuickBird images both in open areas and build-up urban
environments. The results demonstrate its robustness and
viability on extracting salient main road centerlines.
1057 Forest and Land Cover Mapping in a Tropical Highland Region
Christian Tottrup
Abstract Download
Full Article
Linear mixture modeling, in combination with a set of
decision rules, is used to map tropical forest and land-cover
classes within a highland region in north central Vietnam.
The linear mixture model is applied to a topographic
normalized SPOT HRVIR image and the resulting fractional
images of three endmembers: green vegetation, shade, and
soil are classified into six forest and land-cover classes
using field data and a decision tree classifier. The overall
accuracy of the classification is assessed to 82.1 percent
(KHAT 0.78), which is significantly better than a decision
tree classification based on the four SPOT HRVIR reflectance
bands. In conclusion, the presented classification approach
is believed to advance the use of satellite remote sensing in
support for land-use planning and natural resource management
in areas of complex terrain.
1067 Integration of Ikonos and QuickBird Imagery for Geopositioning Accuracy Analysis
Rongxing Li, Feng Zhou, Xutong Niu, and Kaichang Di
Abstract Download
Full Article
This research investigated the accuracy in three-dimensional
(3D) geopositioning achieved by integrating Ikonos and
QuickBird images using the vendor-provided rational
polynomial coefficients (RPCs). One pair of stereo Ikonos
images and one pair of stereo QuickBird images were
collected for the same region of Tampa Bay, Florida, and
used in this study. Results of 3D geopositioning from different
combinations of Ikonos and QuickBird stereo images were
generated by using an improved rational function model
(RFM). The relationship between the satellite-borne pointing
geometry and the attainable ground accuracy is examined.
This research demonstrates that the integration of Ikonos
and QuickBird images is feasible and can improve the
3D geopositioning accuracy using a proper combination
of images.