Peer-Reviewed Articles
1365 Probing the Relationship Between
Classification Error and Class Similarity
Ola Ahlqvist and Mark Gahegan
Abstract Download Full Article
We present a rationale and method for representing the
vagueness in taxonomic class definitions in cases where
classes are described by a set of characteristics, such as
those sometimes used as the basis for land-cover category
discrimination. We further describe methods to estimate the
semantic similarity between any two classes by calculating
semantic similitude metrics based on such parameterized
class definitions. Our working hypothesis is that a large
similitude would predict categories that will be more prone
to confusion and hence image or map misclassification.
We use two different existing data sets to demonstrate and
evaluate the method, and the results support our original
hypothesis. Consequently, we argue that classification
schemes that are based on parameterized definitions could
be assessed for problematic categories during their construction using our approach, and thus, enabling the identification of a thematic vagueness component to supplement the
more traditional statistical measures derived from the error
matrix.
1375 Automatic Camera Placement in Vision
Metrology Based On A Fuzzy Inference System
Mohammad Saadatseresht, Farhad Samadzadegan, and Ali Azizi
Abstract Download Full Article
Automatic determination of the camera placement for
measuring complex industrial objects by vision metrology
systems is a real challenge if the 3D simulated CAD models
of the object along with the workspace information are not
available. In such a case, several uncertain parameters
such as visibility, accessibility, and camera-object distance
are introduced into the camera placement decision-making
that make it a non-deterministic complicated process.
Hence, the uncertain behavior of the vision constraints
demands the use of the fuzzy logic inference approach for
the camera placement network design. In this paper, a
novel method based on fuzzy logic reasoning strategy is
proposed for the accuracy enhancement of an existing
photogrammetric network by automatically adding new
exposures. The fuzzy system is designed to make use of
human type reasoning strategy by incorporating appropriate rules. The results indicate the high potential of the
proposed method for automatic sensor placement in vision
metrology.
1387 An Evaluation of Remote Sensing-derived
Landscape Ecology Metrics for Reservoir
Shoreline Environmental Monitoring
Mark W. Jackson and John R. Jensen
Abstract Download Full Article
The goal of this study was to determine the value of including landscape ecology patterns and structure metrics
extracted from high-resolution, remotely-sensed imagery in
the development of a Shoreline Environmental Impact Index
(SEII). Methods of combining landscape ecology metrics to
create a meaningful Shoreline Environmental Impact Index
included multiple linear regression, multiple discriminant
analysis, genetic neural networks, and feed-forward, back-propagation neural networks. The landscape ratings produced by the SEII’s generated using these methods were then
compared to landscape ratings by experts. There was very
little difference in the performance of several SEII’s generated
despite differences in metrics and their weighting chosen by
the different methods. The ratings from all methods showed
their ability to reflect the expert ratings with moderate
accuracy: ≤84 percent agreement. Conclusions indicate that
the contributions of landscape metrics to the ability of an
SEII to discriminate between levels of shoreline degradation
are variable, dependent upon the method of combination.
Any of the current forms of the SEII is suitable for generating
general indication of shoreline health.
1399 Automatic Segmentation of High-resolution
Satellite Imagery by Integrating
Texture, Intensity, and Color Features
Xiangyun Hu, C. Vincent Tao, and Björn Prenzel
Abstract Download Full Article
High spatial resolution satellite imagery has become an
important source of information for geospatial applications.
Automatic segmentation of high-resolution satellite imagery
is useful for obtaining more timely and accurate information. In this paper, we develop a method and algorithmic
framework for automatically segmenting imagery into different regions corresponding to various features of texture,
intensity, and color. The central rationale of the method
is that information from the three feature channels are
adaptively estimated and integrated into a split-merge
plus pixel-wise refinement framework. In the procedure
for split-merge and refinement, segmentation is realized
by comparing similarities between different features of
sub-regions. The similarity measure is based on feature
distributions. Without a priori knowledge of image content,
the image can be segmented into different regions that
frequently correspond to different land-use or other objects.
Experimental results indicate that the method performs
much better in terms of correctness and adaptation than
using single feature or multiple features, but with constant
weight for each feature. The method can potentially be
applied within a broad range of image segmentation
contexts.
1407 Acquisition of Through-water Aerial Survey
Images: Surface Effects and the Prediction of
Sun Glitter and Subsurface Illumination
Richard Mount
Abstract Download Full Article
The behavior of light at the air/water interface has substantial effects on the quality of vertical, or nadir-looking imagery
used to interpret subsurface features for purposes such as
marine habitat mapping. Reflection of the direct solar beam
into the sensor by waves on the surface of the water creates
bright glints, which obscure bottom features of interest. Sun
angle, refraction, and reflection of the direct solar beam
affect the amount of subsurface illumination and shadowing
of bottom features. Simple interpretations of these sea
surface effects are made with sufficient accuracy to improve
planning for airborne, vertical image capture, particularly
aerial photography or video imagery. The time available for
image capture over shallow water is typically limited to a
short period in the morning. The start time is controlled by
subsurface illumination levels, which are determined by sun
angle and locally variable factors, such as light attenuation
by the water column, rather than surface reflection or
subsurface shadowing. The end time is determined by sun
glitter effects, which in this case study, are predictable from
sun angle, camera field of view, and wind speed with an R2
value of 0.9554.
1417 Derivative Analysis of AVIRIS Data
for Crop Stress Detection
Lee Estep and Gregory A. Carter
Abstract Download Full Article
Low-altitude Airborne Visible/Infrared Imaging Spectrometer
(AVIRIS) hyperspectral imagery of a cornfield in Nebraska
was used to determine whether derivative analysis methods
provided enhanced plant stress detection compared with
narrow-band ratios. The field was divided into 20 plots
representing four replicates each of five nitrogen (N) fertilization treatments that ranged from 0 to 200 kg N/ha in 50 kg/ha
increments. The imagery yielded a 3 m ground pixel size for
224 spectral bands. Derivative analysis provided no advantage in stress detection compared with the performance of
narrow-band ratio indices derived from the literature. This
result was attributed to a high leaf area index at the time
of the overflight (LAI of approximately 5 to 6) and the high
signal-to-noise character of the narrow AVIRIS bands.
1423 A Methodology for Spatial Uncertainty
Analysis Of Remote Sensing and GIS Products
Guangxing Wang, George Z. Gertner, Shoufan Fang, and Alan B. Anderson
Abstract Download Full Article
When remote sensing and GIS products are generated,
errors and uncertainties from collection, processing and
analysis of image and ground data, and model development, accumulate and are propagated to the maps. The
products thus possess many sources of uncertainties that
vary spatially and temporally. Spatially identifying the
sources of uncertainties, modeling their accumulation and
propagation, and finally, quantifying them will be critical
to control the quality of spatial data. This paper demonstrates a methodology and its applications for a case study
in which uncertainty of predicted soil erosion is hierarchically partitioned into various primary components on
a pixel-by-pixel basis. The methodology is based on a
regionalized variable theory of variables. It integrates
remote sensing aided co-simulation algorithms in geostatistics, and uncertainty and error budget methods in
uncertainty analysis. The simulation algorithms generate
realizations that can be used to calculate local estimates,
and the variances and co-variances between them. Uncertainty and error budget methods partition the uncertainty
of output into various input components and quantify
their relative uncertainty contributions. The results can
thus suggest the main uncertainty sources and their variation spatially, and further provide a rationale to reduce
errors in map generation and application.