QUANTITATIVE ASSESSMENT OF SHORELINE CHANGES
USING
MULTI-TEMPORAL SATELLITE IMAGES
Liang-Chien Chen
Professor
and Jiann-Yeou Rau
Assistant Research Scientist
Center for Space and Remote Sensing Research
National Central University, Chung-Li, Taiwan 32054, R. O. C.
E-mail: lcchen(gcsrsr.ncu.edu.tw
Commission VII, Working Group 8
KEY WORDS: Multi-Temporal Satellite Images, Surface Reconstruction , Coast.
ABSTRACT:
We present here a new scheme that performs surface reconstruction and coastline change detection for tideland
areas using multi-temporal satellite images and tidal measurements. The basic idea of this investigation is to
reconstruct a reference DTM for a tideland area of interest from a set of multi-temporal SPOT images and tidal
measurements first. Then, the coastlines on a historical target satellite images are extracted and compared to the one
predicted from the reference DTM according to the associated tide elevations. The variations of shorelines between
the reference image set and the historical images may thus be assessed quantitatively. À self-consistency check may
also be performed to evaluated the quality for predicted shorelines. Experimental results indicate that the relative
error is smaller than 5% as far as the area variation is concerned. The largest sand bar of Taiwan ,Wai-San-Ting, has
lost its area about 37% during the last eight years.
1. INTRODUCTION
Detection of landcover changes for coastal areas is
an important task in environmental monitoring
[Hull ,1978]. The variations of shorelines have direct
impacts on the economy development, land planning ,
and land management. Thus, The terrain changes of
tideland have attracted world-wide interests.[Welch et
8l.,1992 ; Stokkom et al., 1993]
Carter[1978] investigated the applicability of satellite
images in the data acquisition for wetlands. The
spatial resolution of the satellite images i.e., Landsat
MSS, is limited to 80M at the time. Frihy et al., [1994]
identified the pattern of shoreline changes for Nile
Delta. However, the tidal variation is not rigorously
treated. In addition, three dimensional terrain analysis
for shore areas is no considered.
The spatial, spectral , and temporal resolutions of
satellite images have significantly improved in recent
years. Hence, combining multi-temporal satellite
images and tidal measurements , through image
interpretation and analysis, make the detection of
dynamic changes for shorelines possible. The basic
idea of this investigation is to reconstruct a reference
DTM for a tideland area of interest from a set of multi-
temporal SPOT images and tidal measurements first.
Then, the coastlines on a historical target satellite
image are extracted and compared to the one traced
from the reference DTM according to the associated
tide elevations. The variations of shorelines may thus
be assessed quantitatively.
The structure of this paper is as follows. In the next
section ,we give a brief introduction to the techniques
118
for the proposed scheme, Section 3 provides the
experimental results. Finally, section 4 offers some
conclusions.
2. THE PROPOSED SCHEME
The proposed scheme comprises two major
components. The first is to derive a reference DTM for
tideland areas from a set of reference satellite images
incorporating different tide elevations. The second
component predicts the shoreline for a given tide
elevation from the reference DTM. Then a change
detection for the shoreline may be performed with
respect to a historical data set. The flow chart of the
proposed scheme is shown in figure 1.
2.1 Generation of a Reference DTM
The procedure for derivation of a reference DTM
includes:
(1)collecting a series of SPOT data as reference
images which are associated with different tide
elevations over a short period of time, when
insignificant variations for a tideland area of interest
are assumed,
(2)performing geometrical registrations among the
reference images,[Chen & Lee, 1992]
(3)Extracting shorelines from each of the reference
images,
(4)assigning elevation for each point on the extracted
shorelines according to tidal measurements,
(5)superimposing all layers of the extracted shorelines
with different elevations,
(6)generating a reference grid DTM through an
interpolation procedure. [Lee & Chen 1990]
International Archives of Photogrammetry and Remote Sensing. Vol. XXXI, Part B7. Vienna 1996