www.aoe.cas.cn
Recent Progress of BaoTou Comprehensive
Cal&Val Site
September 5, 2016
LANDNET WG meeting #4 CEOS/WGCV-41 Plenary
Academy of Opto-Electronics(AOE), Chinese Academy of Sciences National Remote Sensing Center of China (NRSCC), Ministry of Science and Technology , P. R. China
Prof. Chuan-rong Li
报告时间9月5日下午17:00,时长15分钟(包含提问)。
www.aoe.cas.cn 2 2
Outline
(1) • Overview of the Baotou site
(2) • Progress in automatic calibration
(3) • Great events of Baotou site
www.aoe.cas.cn 3
Overview of the Baotou site
• Located in the Inner Mongolia, China; 50km away from the Baotou city.
• A flat area of approximately 300km2, about 1270m above sea level, features a cold semi-arid climate.
• Land covers: Sand, bare soil, grass, lake, various crops ( maize, sunflower, lucern, potato, etc. )
Lake Ulansuhai
3 grassland
Lake Ulansuhai bare soil
sand
sunflower lucern
potato maize
Site overview:
www.aoe.cas.cn 4
Stepwise Cal&Val technical system
Comprehensive surface targets
with artificial targets and natural
scenes reasonably matching
Airborne standard sensors
Ground test equipment
(a) artificial targets
Artificial portable
targets Artificial
permanent
targets
Natural
scenes
Data processing
and quality
evaluation system
Automatic
in-situ
calibration
system
Overview of the Baotou site
• Advanced Cal&Val system
• Great contribution to RS applications
• International platform for R&D
www.aoe.cas.cn 5
Accomplished automated surface spectrum measurement system
Progress in automatic calibration
As one of the four demonstration sites of the Radiometric Calibration Network (RadCalNet),
four automated systems with identical spectroradiometer configurations were set up on the
black, white and gray artificial permanent targets, and the sand field.
CE 318
1.5km
Automated system in sand field
Cimel CE318
Data
processing
server
Automated systems
Worldview3 image over Baotou site on Oct.
27, 2015. (Band 5/3/2)
www.aoe.cas.cn 6
Essential data collection by automated surface spectrum measurement system
Progress in automatic calibration
Essential data of the site has been collected for several years.
Measuring the surface feature with high temporal resolution (2 min) from 8:30 to 16:30
local time, and supporting data self-checking and teletransmission.
0
5
10
15
20
25
30
35
Jul.15 Aug.15 Sep.15 Okt.15 Nov.15 Dez.15 Jan.16 Feb.16 Mrz.16 Apr.16 Mai.16 Jun.16 Jul.16 Aug.16
Black White Gray Sand
Month-year
No
. o
f d
ays w
ith
me
asu
rem
en
t
www.aoe.cas.cn 7
To meet the requirement of RadCalNet portal, the surface reflectance is retrieved from the
observed reflected radiance and the measured atmospheric parameters, with the help of
atmospheric radiative transfer model (MODTRAN);
Automated system only observes surface at fixed point. Because of the heterogeneity of the
target, the reflectance should be corrected according to the historical surface measurement.
Progress in automatic calibration
400 500 600 700 800 900 10000
0.2
0.4
0.6
0.8
Wavelength (nm)
Su
rfac
e re
flec
tan
ce
Measured (SVC)
Retrieved (Auto system)
400 500 600 700 800 900 10000
0.2
0.4
0.6
0.8
Wavelength(nm)
Su
rfac
e re
flec
tan
ce
Measured (SVC)
Retrieved (Auto system)
400 500 600 700 800 900 10000
0.2
0.4
0.6
0.8
Wavelenth(nm)
Su
rfac
e re
flec
tan
ce
Measured (SVC)
Retrieved (Auto system)
400 500 600 700 800 900 10000
0.2
0.4
0.6
0.8
Wavelength (nm)
Su
rfac
e re
flec
tan
ce
Measured (SVC)
Retrieved (Auto system)Black White
Gray Sand
Fixed point (green curves) V.S.
Area average (red curves)
scat
sky
diffuse
sky
direct
solar
g
EEE
L
Direct irradiance Sky irradiance
Observed at-ground
reflected radiance
Calculated with MODTRAN and
atmospheric parameters (AOT, water
vapour, …)
*S
Fix point reflectance
Scaling coefficient
Fix
ed
po
int
reflecta
nce r
etr
ieval
Sp
atia
l sca
ling
effe
ct co
rre
ctio
n
www.aoe.cas.cn 8
84 effective files following RadCalNet portal format have been generated and uploaded.
Each file contains information for one day of one target. In the file, there are 13 records for
observations at different time (9:00-15:00 local time, 2 values/hour).
Progress in automatic calibration
0 096 098 099 101 103 104 107 111 112 113 115 119 127 128 129 130 131 133 175 176 204 210 211 213 214 215 216 217 218 222 2230
0.02
0.04
0.06
0.08
0.1
DOY
Ref
lect
ance
Blue Green Red NIR
0 096 098 099 101 103 104 107 111 112 113 115 119 127 128 129 130 131 133 175 176 204 210 211 213 214 215 216 217 218 222 2230
0.05
0.1
0.15
0.2
0.25
DOY
Ref
lect
ance
Blue Green Red NIR
0 096 098 099 101 103 104 107 111 112 113 115 119 127 128 129 130 131 133 175 176 204 2380
0.2
0.4
0.6
0.8
DOY
Ref
lect
ance
Blue Green Red NIR
Bla
ck ta
rge
t G
ray ta
rge
t W
hite
ta
rge
t
www.aoe.cas.cn 9
BRDF characteristics
Progress in automatic calibration
400 600 800 1000 1200-40
-20
0
20
40
60
Wavelength(nm)
Fanis
tro
py(%
)
-30° -20° -10° 10° 20° 30°
400 600 800 1000 1200-40
-20
0
20
40
60
Wavelength(nm)
Fanis
tro
py(%
)
-30° -20° -10° 10° 20° 30°
400 600 800 1000 1200-40
-20
0
20
40
60
Wavelength(nm)
Fanis
tro
py(%
)
-30° -20° -10° 10° 20° 30°
Black white
gray
400 500 600 700 800 900 1000 1100 1200-20
-10
0
10
20
30
40
Wavelength(nm)
Fa
nis
otr
op
y(%
)
30o
20o
10o
-10o
-20o
-30o
sand
The artificial targets, which are paved by gravels, have much more obvious angular
anisotropy than the sand field.
With Ross-Li model, the BRDF effects can be well described. Therefore, after BRDF
correction, the accuracy of calibration can be improved.
• Black target: 0.24%(<10°);
0.46%(<30°)
• Gray target: 0.67%(<10°);
0.74%(<30°)
• White target: 0.95%(<10°);
2.03%(<30°)
• Sand target: 0.478%(<10°);
0.692%(<30°)
RMSE of BRDF model
With BRDF correction, the
angular effects may be reduced
below 1% in most cases.
www.aoe.cas.cn 10
Progress in automatic calibration
The spectroradiometers used in the automated surface spectrum
measurement system were calibrated in NIM (National Institute
of Metrology, China).
The overall uncertainty is reported to be lower than 2%. (k=1,
traced to NIM).
400 500 600 700 800 900 10000
0.5
1
1.5
2
2.5
Wave length (nm)
Sta
bili
ty (
%)
Black
Gray
White
SandBlack: 0.13%
Gray: 0.12%
White: 0.11%
Sand: 0.1%
400 500 600 700 800 900 10001
1.5
2
2.5
3
Wave length (nm)
Co
mb
ine
d U
nce
rta
inty
(%
)
Black
Gray
White
Sand
Overall uncertainty of 4 spectroradiometers
Wavelength/nm 300 400 500 600 700 800 900 1000
Repeatability/% 1 0.7 0.5 0.5 0.5 0.5 0.6 0.8
Lamp/% 0.8 0.6 0.4 0.4 0.4 0.4 0.5 0.6
Current % 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1
Alignment for lamp/% 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
Distance % 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
Diffuser pannel/% 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
Uniformity /% 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
Alignment for
spectrometer/% 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
nonlinearity % 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5
Stray light /% 0.8 0.7 0.5 0.5 0.5 0.5 0.5 0.6
Combined U/%(k=1) 1.84 1.56 1.32 1.32 1.32 1.32 1.4 1.57
Expanded U/% (k=2) 3.7 3.1 2.6 2.6 2.6 2.6 2.8 3.1
Uncertainty budget for lamp-panel system (NIM)
Stability of 4 spectroradiometers
www.aoe.cas.cn 11
Demonstration on satellite calibration
Since the end of February 2016, when all the 4 systems at artificial targets and sand field were
updated with same spectroradiometers,there have been 10 matchups for Landsat8, Sentinel-
2a, ZY3, ZY-02C, GF-1, GF-2 and CBERS-4 satellites.
GF1 (Pan:2m / MUX:8m)
Acquired date: 2016-06-07
GF2 (Pan:1m / MUX:4m)
Acquired date: 2016-04-22
ZY02C (Pan:5m / MUX:10m)
Acquired date: 2016-06-07
CEBRS04 (Pan:5m / MUX:10m)
Acquired date: 2016-06-07
ZY3 (Pan:2.1m / MUX:6m)
Acquired date: 2016-04-20 Landsat 8 (Pan:15m / MUX:30m)
Acquired date: 2016-06-24
Satellite Date
Landsat8 04/21,05/07,06/08,06/
24
07/26,08/04,08/27
Sentinel-
2a
03/30, 04/29, 07/21
CBERS-4 06/07
GF-1 04/23,06/07,08/20
GF-2 04/22
ZY3 04/20
ZY3-02 07/20
ZY-02C 04/16,04/19,04/22,04/
25,04/28,05/12,
06/07
www.aoe.cas.cn 12
Demonstration on satellite calibration
Demonstration satellite in RadCalNet framework
Average differences between simulated TOA radiance and observed TOA
radiance are 4.8%, 1.5%, 0.0%, 1.2% for Blue, Green, Red, NIR band of
OLI/Landsat8.
Blue Green Red NIR0.8
0.9
1.0
1.1
1.2
Channels
Sim
ula
ted
/Ob
serv
ed
2016-04-21
2016-05-07
2016-06-08
2016-06-24
Clear day but large AOD (0.6389@550nm)
2016/04/21 2016/05/07 2016/06/08 2016/06/240
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Date
Atm
osp
heri
c p
ara
mete
rs
Aerosol optical thickness@550nm
Columnar water vapor(g/cm2)
Except for 8 Jun, AODs at
other days are smaller than
0.3@550nm
OLI/Landsat8 image
www.aoe.cas.cn 13
Demonstration on satellite calibration
Chinese high-resolution satellite---ZY3-02
Pan+MSS/ZY3-02
Pan/ B1 B2 B3 B4
Wavelength
(um) 0.5~0.8
0.45~
0.52
0.52~
0.59
0.63~
0.69
0.77~0
.89
Pixel size
(m)
2.1(NAD);
3.5
(FWD/B
WC)
6
swath width
(km) 51 51
revisit
period
(±32°,day
s)
5 5
ZY3-02 satellite was launched on May 30, 2016, which is a new civilian high-resolution
mapping satellite. It will join its predecessor ZY3-01 satellite (launched in 2012) in the same
orbit to form a network and capture high-definition, 3D images and multispectral data.
The on-orbit test was carried out over the Baotou site during July 15-25, 2016.
www.aoe.cas.cn 14
• Measurement Synchronous measurement: spectral reflectance for 3 permanent artificial targets and 6
portable targets; Automatic measurement: spectral radiance for 3 permanent artificial targets and the sand
field; atmospheric parameters.
Demonstration on satellite calibration
Chinese high-resolution satellite---ZY3-02
portable gray-scale targets Portable colored targets
AOD atmospheric profiles sunphotometer atmospheric sounding
synchronous spectrum measurement
www.aoe.cas.cn 15
Demonstration on satellite calibration
Chinese high-resolution satellite---ZY3-02 Chinese high-resolution satellite---ZY3-02
Radiometric calibration
• The portable target of 60% reflectance shows saturation in
the blue band image, so the corresponding point was
removed when fitting calibration coefficient.
• Using 2 portable colored targets to validate calibration result
based on 4 automated measurement systems, as well as
calibration result based on all targets’ synchronous
measurement. Further Improvement is still necessary for
automated calibration mode.
The response linearity were greater than 0.98 for all bands . Validation results by using portal blue colored targets
For saturation, the DN
value of red region is
filled with 977 by
satellite agency
Automatic measurements/
Synchronous measurements
Targets
reflectance
Simulated
targets TOA
radiance
Observed
targets
DN
calibration coefficients:
a(automated mode);
b(Synchronous mode)
Sensors
Sensors
Observed
targets DN
Observed TOA
radiance using
coefficients a/b
Simulated
sand TOA
radiance
Validation Calibration
Measurements on
colored targets
Targets
reflectance
www.aoe.cas.cn 16
Demonstration on satellite calibration
Chinese high-resolution satellite---ZY3-02
Calculation Specification
Cross track 0.153
0.1 Along track 0.103
MTF
edge angle SNR MTF-ref Mean of relative
errors(100 times) RMSE
(100 times) Error <10% (100 times)
-8.7610° 42 0.15 -0.83% 20.76% 38
Error analysis results for 100 simulated images
Estimate errors of 100 simulated images
The error analysis indicates that
influenced by factors such as
black/white contrast, atmosphere
and sensor noise, the SNR value
when estimating MTF is only 42,
and in addition the edge angle is
larger than the recommended
threshold 5°, so the current MTF
evaluation may contain relatively
no small uncertainty. -6,00%
-5,00%
-4,00%
-3,00%
-2,00%
-1,00%
0,00%
1,00%
2,00%
0 1 2 3 4 5 6 7 8 9 10
Re
lati
ve
err
or
Edge angle
Relative estimate errors for different angles
www.aoe.cas.cn 17
In August 2016, the first workshop of the “National Calibration and Validation Site
for High Resolution Remote Sensors” was held at Baotou. Representatives of all
EO satellite operational agencies of China attended the workshop.
Great Events
www.aoe.cas.cn 18
The 28th IVOS meeting and the 5th RadCalNet meeting were hosted by Baotou site during July
18 to 21 2016. Nearly 30 worldwide scientists from NASA, NOAA, ESA, CNES, DLR, NPL,
CSIRO, ONERA, GSJ, SDSU, etc, attended this meeting.
Great Events
www.aoe.cas.cn 19
After this meeting, on July 22, four scientists (Dr. Steffen Dransfeld from ESA, Dr.
Françoise Viallefont-Robinet from ONERA, Dr. Esad Micijevic from USGS and Dr.
Hirkazu Yamamoto from GSJ) visited the Baotou site.
Great Events
www.aoe.cas.cn
20
Future work
To improve the operational capability
Robust our system in operational running, especially the improvement of controlling
and data processing software.
Extension of spectrum range to SWIR domain.
To improve the calibration accuracy
Development of the local atmospheric model:Explore the information from historical data to find the characteristics of the local atmosphere and surface. Study how to improve or correct the atmospheric radiative transfer model.
Development of automated measurement device for total/diffused sky irradiance, providing observed total irradiance to replace the simulation value derived from MODTRAN in current system.
www.aoe.cas.cn 21
To Enhance the capability of land product validation based on the infrastructure of Baotou Site
crop (maize, wheat, sunflower, melon)
Future work
Install evaporation, soil moisture observation equipment of Eddy covariance system (EC), Large aperture scintillometer (LAS) and TDRs, to realize the supporting capacity of land surface product modeling and validation;
Study on the temporal, spatial, spectral and viewing angle matching technologies, decreasing significant scaling bias of step-wised validation.
Eddy covariance system (EC)
Large aperture scintillometer (LAS)
MiniTrase COSMOS
Detector ESU of 30m×30m
Soil moisture profile
The distributed
soil moisture
monitoring
system
4D observation
platform
The matrix evaporation
observation system
grass
www.aoe.cas.cn 22
Activities on Baotou site