Peer-Reviewed Articles
897 The Role of Soft Classification Techniques in
the Refinement of Estimates of Ground Control Point Location
Giles M. Foody
Abstract Download
Full Article
Mis-registration of data sets is one of the largest sources of error in many
remote sensing studies. An initial contribution to this error arises through
the mis-location of ground control points (GCPs) used to derive geometrical
transformation equations. Here, it is proposed that a soft classification
of land cover may be used to direct the estimation of GCP location. The soft
classification provides an estimate of the class composition of each image
pixel. The spatial distribution of a pixel's component land covers may then
be modeled over the area it represents and used to reduce the error in estimating
the location of a GCP that lies within this area. An example is provided
in which the error in locating a set of GCPs was reduced by up to 35.7 percent
when information from a soft classification was available to aid the estimation
of their location at a sub-pixel scale.
905 A Comparison of Fuzzy vs. Augmented-ISODATA
Classification Algorithms for Cloud-Shadow Discrimination from Landsat
Images
Assefa M. Melesse and Jonathan D. Jordan
Abstract Download
Full Article
Satellite images are the most important source of land-cover data over a large
range of temporal and spatial scales. However, the complete realization of
satellite imagery as a source of land-cover information is limited by the
presence of contaminants such as cloud and associated shadows in the image.
These contaminants are not adequately handled with conventional image classification
techniques such as the unsupervised maximum-likelihood technique. This study
comprises a comparison of two classification algorithms, the fuzzy technique
and an augmented form of the Iterative Self-Organizing Data Analysis (ISODATA)
technique, which were used to discriminate low-altitude clouds and their
shadows on a Landsat Thematic Mapper (TM) image of the Econlockhatchee River
basin (Econ), in central Florida. Preliminary conventional unsupervised maximum-likelihood
classification of the image resulted in clouds being mixed with built-ups
and shadows being mixed with water bodies. Regions containing these two kinds
of mixed categories were first masked, then fuzzy and augmented-ISODATA classifications
were performed on them. The ISODATA classification algorithm was run on the
TM visible/shortwave bands and augmented with scatter diagrams of surface
temperature versus several vegetation indices; the fuzzy algorithm was run
on TM bands 1 through 5 and band 7. An accuracy assessment of the techniques
was carried out using 40 randomly selected points within the image. Results
of the classifications showed that both algorithms successfully discriminated
clouds from other bright features, and shadows from other dark features,
with an overall accuracy of greater than 80 percent.
913 Techniques for Mapping Suburban Sprawl
Jeanne Epstein, Karen Payne, and Elizabeth Kramer
Abstract Download
Full Article
The pervasive problems generated by urban sprawl have prompted us
to examine methods for delineating the extent of suburban land
cover in Georgia. This
paper assesses the advantages and disadvantages of two different methods
of mapping suburban neighborhoods: traditional unsupervised classification
of
Landsat 5 TM data and a newly devised procedure for editing and buffering
road coverages. We conclude that, while the amount of time required
to edit and
buffer road coverages is significantly higher than that for traditional remote
sensing techniques, the improved thematic accuracy, spatial contiguity, and
potential future uses of the resulting dataset justifies its use in a state-wide
mapping program.
919 Improvement of an Oak Canopy Model Extracted
from Digital Photogrammetry
Peng Gong, Xueliang Mei, Gregory S. Biging, and Zuxun Zhang
Abstract Download
Full Article
Digital surface models (DSMs) automatically derived with digital photogrammetric
systems are useful in land surface change monitoring, including forest growth
monitoring. However, they cannot be applied directly to forest canopy change
analysis with high accuracy due to the inevitable deficiencies in existing
commercial digital photogrammetry packages. In a hardwood rangeland monitoring
study, we found that the oak tree and woodland canopy boundaries were not
well determined using several digital photogrammetry packages available to
us. There was a noticeable discrepancy between the true crown closure and
that determined by subtracting the DSM and the corresponding DEM that excludes
tree heights. In this paper, we present a correction method for improvement
at the erroneous canopy boundary locations in the DSM using shadow and boundary
information extracted from imagery. The method is designed for correcting
errors for broadleaf tree canopies. Aerial photographs taken over oak woodland
hills were tested. Using manual photogrammetric measurements as the reference,
we found that most of the points (88.3 percent) on the canopy boundaries
were displaced by greater than 1 meter with a conventional digital photogrammetric
package. After the proposed algorithms were applied, greater than 98.6 percent
of the points on canopy boundaries were found to be within 1 meter of their
reference positions. 78.4 percent of the reference points had greater than
2 meters elevation errors with the conventional package while greater than
85.6 percent of those points were found to be within 2 meters of the eference
after the proposed algorithms were applied.
925 Detecting and Measuring Individual Trees Using
an Airborne Laser Scanner
Åsa Persson, Johan Holmgren, and Ulf Söderman
Abstract Download
Full Article
High-resolution airborne laser scanner data offer the possibility to detect
and measure individual trees. In this study, an algorithm which estimated
position, height, and crown diameter of individual trees was validated with
field measurements. Because all the trees in this study were measured on
the ground with high accuracy, their positions could be linked with laser
measurements, making validation on an individual tree basis possible. In
total, 71 percent of the trees were correctly detected using laser scanner
data. Because a large portion of the undetected trees had a small stem diameter,
91 percent of the total stem volume was detected. Height and crown diameter
of detected trees could be estimated with a root-mean-square error (RMSE)
of 0.63 m and 0.61 m, respectively. Stem diameter was estimated, using laser
measured tree height and crown diameter, with an RMSE of 3.8 cm. Different
laser beam diameters (0.26 to 3.68 m) were also tested, the smallest beam
size showing a better detection rate in dense forest. However, estimates
of tree height and crown diameter were not affected much by different beam
size.
933 Methods for Measuring Height and Planimetry
Discrepancies in Airborne Laserscanner Data
Hans-Gerd Maas
Abstract Download
Full Article
Airborne laserscanning (or lidar) has become a very important technique for
the acquisition of digital terrain model data. Beyond this, the technique
is increasingly being used for the acquisition of point clouds for 3D modeling
of a wide range of objects, such as buildings, vegetation, or electrical
power lines. As an active technique, airborne laserscanning offers a high
reliability even over terrain with poor image contrast. The precision of
the technique is often specified to be on the order of one to two decimeters.
By reason of its primary use in digital terrain modeling, examinations of
the precision potential of airborne laserscanning have so far been concentrated
on the height precision. With the use of the technique for general 3D reconstruction
tasks and the increasing resolution of laserscanner systems, the planimetric
precision of laserscanner point clouds becomes an important issue. In addition
to errors in the laser distance meter and the deflecting mirror system, the
error budget of airborne laserscanning instruments is strongly influenced
by the GPS/INS systems used for sensor pose (position and orientation) determination.
Errors of these systems often lead to the deformation of laserscanner data
strips and may become evident as discrepancies in the overlap region between
neighboring strips in a block of laserscanner data. The paper presents least-squares
matching implemented on a TIN structure as a general tool for the determination
of laser-scanner strip discrepancies in all three coordinate directions,
using both height and reflectance data. Practical problems of applying matching
techniques to 2.5D laserscanner point clouds are discussed and solved, and
the success of the technique is shown on the basis of several datasets. Applying
least-squares matching techniques to dense laserscanner data in a TIN structure,
strip discrepancies can be determined with centimeter precision for the height
coordinate and decimeter precision for the planimetric coordinates.
941 Cloud Mapping from the Ground: Use of Photogrammetric
Methods
Gabriela Seiz, Emmanuel P. Baltsavias, and Armin Gruen
Abstract Download
Full Article
Within the European Union (EU) project Cloudmap, a ground-based sky imager
system consisting of two commercial digital CCD cameras with wide-angle lenses
has been established which can theoretically be used to derive various macroscopic
cloud parameters (cloud-base height, cloud-base wind, cloud amount). In this
paper, we present the method to calculate a digital surface model (DSM) of
the cloud base. It includes both the precise determination of the interior
and exterior orientation of the cameras as well as the automatic derivation
of the cloud-base heights using modern photogrammetric algorithms. The presented
measurements were taken during the Mesoscale Alpine Programme (MAP) in Switzerland
in October 1999. The results from our own matching software and from commercial
photogrammetric systems were validated with semi-automatically measured points
and compared with visual observations, lidar, and radiosonde data from the
MAP Composite Observing Network and satellite-based cloud-top heights from
ERS2-ATSR2. The potential of the system to provide very accurate areal cloud-base
height data was shown. This is important for the objectives of the EU project
Cloudmap2, where it is planned to assimilate various cloud parameters, including
cloud-base height, into cloud and high-resolution numerical weather prediction
(NWP) models.
| Top | Home |