Multi-Parameter SIR-C/X-SAR for Geoscience Study in
China
Prof. Guo Huadong
Institute of Remote Sensing
Applications Chinese Academy of
Sciences
P.O. Box 9718, Beijing 100101
China
OBJECTIVES
To establish the backscatter models for the typical
targets at the land surface and study the penetration phenomena.
To develop the techniques for multi-frequency and
multipolarization SAR image processing and specific geoscience
information extraction.
To use SAR imagery to detect the geological structure
and lithology both in arid and subtropic regions, and study the
archaeology and other geoscientific fields.
To develop the interferometric and polarimetric SAR
data analysis methods and evaluate their roles in the geoscience
study.
PROGRESS
Spaceborne-Airborne-Ground-based Radar Remote
Sensing Experiment
During the first SIR-C/X-SAR mission (SRL-1), the
Chinese airborne SAR (CASSAR) campaign was performed in Beijing
test site simultaneously when SRL-1 overflew the test site. Meanwhile,
the backscattering measurement was carried out on the ground using
the truck-mounted scatterometer, and ground-truth data were collected
in real time. This campaign is called the spaceborne airborne
ground-based radar remote sensing experiment (Figure1).
The CASSAR operated in X-band and HH, VV polarizations
with the resolution of 10m°¡10m. The flight altitude
was 7200m. The azimuth and range directions were designed to coincide
with the SRL1. The incidence angle is 73.93°Ñ°83.12°Ñ.
The truck-mounted scatterometer operated in X and L-band with
4 polarizations. The antenna was fixed in the boom about 12m high.
It could illuminate any azimuth directions and elevation angles
from 0°Ñto 84°Ñwith 6°Ñinterval.
The Lunenburg sphere was used for external calibration before
the measurement. The measured distributed targets including bare
soil, winter wheat, rice staple, water body, etc. Meanwhile, vegetation
and soil parameters were measured in the real time, such as biomass,
soil moisture and surface roughness.
The backscattering data acquired in the platforms
with different altitudes could be used for calibration and better
understanding of interaction mechanisms of electromagnetic radiation
with natural surfaces. Then the soil moisture and water content
of winter wheat were analyzed using CASSAR data. The calibration
results and ground truth data analysis of this campaign are put
forward.
Penetration Experiment
The Inner Mongolia test site, which is located in
the east of Badanjilin desert, was selected for the penetration
phenomena experiment, where the vegetation coverage is quite sparse,
and the land surface is mainly covered by the Gobi and sand. The
weather is very dry. Before the SIR-C/X-SAR mission, twelve corner
reflectors were buried in the sand, in which four are under the
natural sand, six in the man-carried dry sand and the last two
on the land surface.
The area has two test sites. No.1 site is located
in western Alashan Trigonum, which is low-lying land composed
of rocks covered by the sand of 1m thickness, with some rocks
coming off the surface occasionally. Four corner reflectors are
buried here, whose buried depth is 1m. Its purposes are (1) to
observe the penetration phenomena of microwave in natural condition
and the media moisture effects on microwave attenuation; (2) to
expose the an process of microwave in natural sand. No. 2 site
was choosed in a Gobi area, where the sand layer of 20cm thickness
covered on the alluvium surface. Six corner reflectors were buried
in a digged place with 0.93 2.73m depth covered by man-carried
sand. Its purpose is to study the attenuation process of microwave
in the dry sand and measure the penetrating height and condition.
Then the samples were collected for measuring the moisture and
complex dielectric constants. The penetration depths through the
sand sheet of multifrequency radar signals were calculated and
explained.
Targets Discrimination and Detection
a.At the Kunlun site of western China, a joint field investigation was carried out with JPL. The volcanoes northeast of Aksayqin Lake, in the Kunlun Mountains of western China, have been studied by combining field observations and SIR-C/X-SAR images. From the study, the volcanic morphology has been identified, which can be separated from the surrounding bedrock and alluvium (figure 2). The effects of different band and polarization combinations in recognizing volcanic features, lava flows, alluvial fans, and bedrock. A number of rock samples collected from the lava flows were analyzed with purpose of measuring dielectric constant isotopic dating, and determining chemical compositions of the samples.
b.Zhao Qing test site in southern China is located
in the subtropical region, where the application of remote sensing
in geological survey and mapping is limited by the dense vegetation
and soil cover and poor data availability due to clouds and rain.
So it was selected to demonstrate the advantages of multifrequency
and multipolarization SAR technology in detecting the geological
structures and lithology underneath a vegetation canopy to evaluate
the capacities and limitation of SAR system for geological mapping
in a subtropical region. The study presents the preliminary interpretation
of geological structures, beds, and lithostratigraphic units from
the shuttle imaging radar data in the test site(figure 3).
c.At the Yanchi site of north-central China, the
study for the ancient segment of the Great Wall of China was done
using the SIR-C/X-SAR image. The Great Wall has been eroded and
buried in place by centuries of blowing sand.
INSAR for DEM Generation
This research used SIR-C/X-SAR L-band data in Xinjing
for extraction of 3D information. Xinjiang is located in the north
of Tibet Plateau, Northwest China. The population is sparse and
the tectonic activity is frequent in the test site. It is the
best site for the study of INSAR for crust deformation, which
is one of the supersites for SIR C/X-SAR. Meanwhile, the vegetation
coverage is less there, which is suitable for INSAR research.
The procedure of INSAR for DEM includes INSAR signal
data processing, normalization of imaging parameters, precise
registration of images, phase correction for flat terrain and
phase unwrapping. After the above processing procedure, slant-range
DEM could be obtained. The ortho projected DEM could be calculated
using imaging parameters.
SIGNIFICANT RESULTS
During the first SIR-C/X-SAR mission, the truck-
mounted scatterometer data acquired in the real time were used
for calibration of the Chinese airborne CASSAR data and SRL-1
survey images in the Beijing test site. In the image linear range,
the regression precision is less than 2 dB (except L-band data).
However, due to the limited dynamic linear range of distributed
targets, it is better to use active calibrators to extend the
dynamic range if calibration. The analysis shows that the correlation
is higher between soil moisture and SAR image than between the
water content of winter wheat.
The penetration experiment indicates the depth of
SIR- C/X SAR penetrating the dry sand in Alashan area is 2.82m,
but some objects in the place more than 2.82m depth can still
be monitored. The limit detecting depth can be calculated from
the correlation between penetration depth and radar response intensity.
L- band can detect the bedrock under the sand sheet. The research
are (1) the depth of sand sheet is 1m thus the bedrock within
0.65m can be detected. (2) less than 1/e energy, which is backscattered
by both the sand layer and rock, can reach the radar antenna.
Therefor, when the water content is less than 6% without great
depth, the bedrock under the sand can be detected.
The volcanic terrain, including in nine cinder cone
and two types of lava flows (pahoehoe and aa lavas), can be identified
from SIR-C/X-SAR images and field observations(figure 4). On the
basis of analyses of backscatter coefficients, LHV appears best
for discriminating the two types lava flows and alluvium and bedrock.
This indicates that depolarization is important and that the roughness
scales important for discrimination of the surfaces are nearer
the scale of L-band (25cm) than the shorter wavelengths. CHV exhibits
a similar effect to that of LHH for separating different surface
features, which may be the result of some depolarization at the
C-band scale (5cm) as well. In general, the shorter wavelengths
of Cand X-band are poorer for delineating the surfaces mapped
in this study. Geological study has shown that the volcanoes in
this area are high in K2O and were erupted between 7.45 and 3.97
Ma. The ages suggest at least two volcanic eruption phases. The
compositions of the volcanoes are associated with collisional
tectonics.
SIR-C/X-SAR data for geological applications in the
Zhao Qing test site highlight the advantages of the multifrequency
and multipolarization data sets in detecting geological structures
and discriminating lithology. The advantages are summarized as
follows: (1) the high resolution, to reveal the geological features
and details underneath the vegetation canopy; (2) multiparameters,
to produce color composite images containing both the spectral
and polarization information, improving the interpretability in
geological applications; and (3) wide swath and less variations
of distortion: to provide a large imaging area to trace and compare
structures and their evolution. The three advantages of the SIR-C/X-SAR
data indicate its potentials in the geological mapping, mineral
exporation, lithlogy and rock type discrimination, and regional
structural study. The major achievements of this study are as
followings: (1) modification and updating of regional structures
and their relationships; (2) discovery of unknown faults and lineaments
which are not reported in the geological literature to date; (3)
reveal the stratification and triangular facets of formation which
indicate the occurrence of bedded structures underneath the vegetation
canopy. This suggests SAR data have great potential in geological
applications. Especially in a moist vegetated environment, it
has incomparable advantages over visible and infrared remote sensing
technology.
In SIR-C/X-SAR images, two generations of the Great
Wall of China can be recognized(figure 5). The most recent version
of the wall was built by the Ming Dynasty during the 14th century.
An older version, built during the Sui Dynasty, runs parallel
to the present wall. Using radar to look at archeological structures
has been very powerful because the radar is sensitive to vertical
structure, such as wall. For these segments of the Great Wall
buried in place by the winds blowing, they can also show up quite
clearly in radar image due to the radarís ability to penetrate
through layers of dry sand. This indicates that the multiple channels
of SIR-C/X-SAR system have the capability to detect different
kinds of structures.
In the phase image and coherence image acquired from
INSAR data (figure 6a), it can be seen that, the phase variance
coincides with the terrain height. The coherence is related to
S/N of data, coredistration precision, look numbers, etc. The
acquired DEM is compared with the 1:100,000 topographic map, which
is only topographic data here. DEM shows much information in details(figure
6b-d).
PUBLICATIONS
Guo Huadong, Zhu Liangpu, Shao Yun and Lu Xinqiao,
1996, Detection of Structural and Lithological
Feature Underneath a Vegetation Canopy Using SIR-C/X
SAR Data in Zhao Qing Test Site of Southern China, Journal
of Geophysical Research, 101(E10):23101-23108.
Guo Huadong, Liao Jingjuan, Wang Changlin, Wang
Chao, Thomas G. Farr and Diane L. Evans, 1997,
Use of Multifrequency and Multipolarization Imaging Radar
for Volcano Mapping in the Kunlun Mountains of Western
China, Remote Sensing of Environment, 59(2), p364-374
Guo Huadong, 1997, Spaceborne Multifrenquency, Polarimetric and
Interferometric Radar for Detection of the Targets on Earth Surface
and Subsurface, Journal of Remote Sening
(Chinese), Vol.1, No.1, p32-39
Wang Chao, Guo Huadong, Li Lin, 1995, Spaceborne
Airborne-Ground based Radar Remote Sensing
Experiment, Chinese Science Bulletin, 40 (24).
Wang Chao, 1997, Using SIR-C Interferometric Data for
DEM Acquisition, Journal of Remote Sening
(Chinese),Vol.1, No.1, p46-49
List of the figures
Figure 1. The sketch of spaceborne-airborne-ground
based remote sensing experiment at Beijing test site of China.
(File name: fig1.jpq)
Figure 2. The color composite image of LHH (red),
LHV (green), and CHH (blue) at Kunlun site of western China. It
appears that the volcanic morpholoy can be separated from the
surrouding bedrock and alluvium. (File name: fig2.tif)
Figure 3. (a)The color composite image of LHH (red),
LHV (green), and CHV (blue) . A group of color stripes in cyan,
greyish brown represents the beds of sedimentary rocks underneath
the vegetation cover. (b) the color composite image of LHH (red),
CHV (green), and XVV (blue) at Zhao Qing test site of southern
China. It shows the structural and lithological features. (c)
Inferred gold-bearing arcuate structure at Gold Pit village. (File
name: fig3.jpg)
Figure 4. Geological interpretation map of the SIR-C/X-SAR
image at Kunlun site of western China. 1. cinder cone and number;
2. aa lava; 3. pahoehoe lava; 4. Triassic strata; 5. Quaternary
alluvium; 6. sample sites. (File name: fig4.tif)
Figure 5. The color composite image of LHH (red),
LHV (green), and CHV (blue) at the Yanchi site of north-central
China. Two generations (Ming and Sui Dynasty) of the Great Wall
can be recognized. (File name: fig5.jpg)
Figure 6. Using SIR-C interferometric data for DEM acquisition at Kunlun site of western China. (a) the phase coherence image; (b) the DEM acquired from L band, VV polarization data; (c) the 3D view of DEM; (d) the 1:100,000 topographic map in the test site. (File name: fig6.jpg)