Peer-Reviewed Articles
1143 Landsat and Apollo: The Forgotten Legacy
Paul D. Lowman, Jr.
Abstract
Download
Full Article
This paper demonstrates that Landsat was fundamentally a result of
the Apollo Program. The U.S. Geological Survey’s EROS proposal
of 1966, which eventually led to Londsat, was stimulated largely by
the demonstrated utility of 1100 orbital photographs from the Gemini
missions, Gemini being solely preparation for Apollo. In addition,
Earth-oriented remote sensing research sponsored by NASA in the mid-1960s,
primarily support for Apollo lunar missions, included studies of Earth
resource applications as well. Finally, the extensive series of airborne
remote sensing studies carried out by the NASA Manned Spacecraft Center
was Apollo-derived in that the primary mission of MSG was to accomplish
a lunar landing. It is concluded that, had it not been for the Apollo
Program, Landsat or its equivalent would have been delayed by 10 years
or more.
1149 Improvement in Predicting Stand
Growth of Pinus radiata (D. Don) across Landscapes Using NOAA
AVHRR and Landsat MSS Imagery Combined with a Forest Growth Process
Model (3-PGS)
Nicholas Coops
Abstract
Download
Full Article
Recent detailed physiological and micro-meteorological studies of forest
ecosystems have lead to new insights that greatly simplify the prediction
of gross primary production (PG) and above-ground net primary production
(NPPA) which are key variables related to conventional measures of
forest growth, such as mean annual increment (MAI) of stem wood.
These simplifications were applied in a monthly time-step model (Physiological
Principles Predicting Growth using Satellite data (3-PGs)) which
requires monthly weather data (average minimum and maximum temperatures
and precipitation), an estimate of soil texture, rooting depth, and
the fraction of photosythetically active radiation absorbed by the
forest canopies (fPAR) which is estimated from a satellite-derived
normalized difference vegetation index (NDVI).
The model was originally tested at sites in Australia and New Zealand using coarse spatial resolution AVHRR Pathfinder data which effectively limited the predictions of NPP^, to broad areas. In this paper, AVHRR and Landsat MSS data are both used by the model, allowing 3-PGS predictions to be applied at a more refined landscape scale.
Accumulated above ground biomass predicted by the model was compared with biomass data from discrete stands in a 2,265-ha Pinus radiata (D. Don) plantation in southern New South Wales, Australia. There was a linear relation between predicted and measured wood production (r2 = 0.84). Additionally, analysis of the results indicated the incorporation of MSS and AVHRR data allowed a variety of stand-specific disturbances to be accounted for, such as thinning.
1157 Development of Landscape Metrics for Characterizing Riparian-Stream
Networks
Michael J. Schuft, Thomas J. Moser, P.J. Wigington, Jr., Don L. Stevens,
Jr., Lynne S. McAllister, Shannen S. Chapman, and Ted L. Ernst
Abstract
Download
Full Article
Sampling methods and functionally related landscape metrics were developed
for characterizing riparian-stream networks using aerial photography
and GIS. A sample area was empirically derived by using morphological
characteristics of increasing portions of the stream network surrounding
points selected on streams. GIS functions were used to band stream
networks in 10-rn increments to a distance of 300m, within which
land cover was interpreted from aerial photographs and digitized.
Incremental banding is an effective approach for characterizing the
composition and pattern of land cover as a function of distance from
the stream network. Structural attributes that capture the linear
nature of riparian-strearn networks, such as the composition, width,
longitudinal extent, and connectivity of woody vegetation, were characterized.
The methods developed provide a flexible framework for deriving landscape
metrics of functionally important structural attributes of riparian-stream
networks for exploring relationships at varying spatial scales with
indicators of stream ecological condition.
1169 Spatial Interrelationships between
Lake Elevations, Water Tables, and Sinkhole Occurrence in Central
Florida: A GIS Approach
Dean Whitman, Timothy Gubbels, and Linda Powell
Abstract
Download
Full Article
Sinkholes constitute the principal geologic hazard in central Florida.
Local hydrogeology is recognized as an important factor in their
formation. We use a GIS to investigate the spatial relationships
between hydrogeology and sinkhole formation near Orlando, Florida.
Landsat TM imagery, digital topography, and well data are used to
construct a model of the head difference between a discontinuous
set of surficial aquifers and the Floridan aquifer, a regionally
extensive confined aquifer. This model is quantitatively compared
to a buffer model of distance to nearest sinkhole constructed from
a database of collapse events. Sinkhole occurrence is positively
associated with regions where the head difference is between 5 and
15m. In these regions, sinkholes are more common and more closely
spaced than expected. In contrast, sinkholes are less frequent and
farther apart than expected in regions of low head difference. This
association of sinkhole proximity to high head difference demonstrates
the importance of hydrostatic loads in sinkhole hazard.
1179 An Image Processing Chain for Land-Cover Classification
Using Multitemporal ERS-1 Data
Jan Verhoeye and Robert De Wulf
Abstract
Download
Full Article
Given the frequent cloud cover, regular updating of land-cover maps
of tropical areas using optical satellite data is problematic. As
weather-independent ERS images are available on a regular basis,
the question is raised as to whether they can be used to produce
a land-cover map.
A processing chain has been developed: it consists of calibration, resampling, filtering, segmentation, principal component transformation, and supervised classification of multi- tern poral radar images. As input to this processing chain, any number of SAR.PRI images or derived texture images can be used. The output consists of a land-cover map and an accuracy assessment. The procedure has been applied to a series of four SAR images, taken over northeast Costa Rica, which yielded a map with an overall accuracy of 76 percent. The high precision with which the large banana plantations can be mapped is most interesting, both for its economic importance and for environmental monitoring.
1187 Remotely Sensed Change Detection Based on Artificial
Neural Networks
Xiaolong Dai and Siamak Khorram
Abstract
Download
Full Article
A new method for remotely sensed change detection based on artificial
neural networks is presented. The algorithm for an automated land-cover
change-detection system was developed and implemented based on the
current neural network techniques for multispectral image classification.
The suitability of application of neural networks in change detection
and its related network design considerations unique to change detection
were first investigated. A neural-network-based change-detection
system using the backpropagation training algorithm was then developed.
The trained four-layered neural network was able to provide complete
categorical information about the nature of changes and detect land-cover
changes with an overall accuracy of 95.6 percent for a four-class
(i.e., 16 change classes) classification scheme. Using the same training
data, a maximum-likelihood supervised classification produced an
accuracy of 86.5 percent. The experimental results using multitemporal
Landsat Thematic Mapper imagely of Wilmington, North Carolina are
provided. Findings of this study demonstrated the potential and advantages
of using neural network in multitemporal change analysis.
1195 A Technique for Spatial Sampling and Error
Reporting for Image Map Bases
Albert K. Chong
Abstract
Download
Full Article
This paper presents a technique for spatial sampling and error reporting
for image base maps. The technique is based on the coverage of an
image map base, the initial estimated accuracy, and the principle
of error propagation to determine a number of check points. The
location of these check points is randomly generated to obtain a
non-biased evaluation of the overall image map base. A spatial
error modeling formula is introduced to estimate the size of error
at various locations over the image map base. Data from a recent
orthoimage accuracy assessment project is used to show the procedure.
The implementation of this technique has resulted in an improvement in the checking procedure of image map base accuracy assessment.
1199 Detection and Location of Objects from Mobile Mapping
Image Sequences by Hopfield Neural Networks
Rongxing Li, Weian Wang, and Hong-Zeng Tseng
Abstract
Download
Full Article
Detection and location of objects is a challenging issue in mobile
mapping data processing for reducing human operations and enhancing
efficiency, considering the vast amount of data and information acquired
by mobile mapping systems. This paper describes research results of
algorithms based on Hopfield neural networks for utility object detection
and location. Specifically, street light poles are modeled in the three-dimensional
(3D) scene domain and detected by the network with neurons formed by
vector edge features from the model and the mobile mapping images.
The established Hopfield neural network is able to detect light poles
at specific locations. It can also be used to detect and locate all
light poles from a mobile mapping sequence, regardless of their positions.
Such automation is particularly important for automatic generation
of special layers in a utility GIS, for example, traffic signs, fire
hydrants, road centerlines, and others. The developed algorithms and
implementation results are described.
| Top | Home |