Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Remote Sensing in Severe Radiation Environments
Ralph Levy*, Kevin P. Hand**, Robert W. Carlson**Winthrop Wadsworth***, Jens Peter Dybwad***
Daniel Berisford**, Didier Keymeulen**, Jason E. Feldman**
•Quant Engineering, ** NASA JPL, *** D&P Ins truments
•Presented at: 2011 Sensors Tech Forum, Boston, M A
Overview• Based on a spectrometer for the delayed NASA Jupiter Europa
Orbiter mission
• Why go to Jupiter - possibilities and problems
• FTIR Spectral Instrumentation
• Analysis and Removal of Radiation Effects
• Other Applications
• Work funded by: NASA JPL contract #1396649Edgewood Chemical and Biological Center DAAD13-03-C-0035 Quant Engineering internal funding
Moons of Jupiter with Liquid Water
Moons of Jupiter with Liquid Water
Measurements and Problems
WAVELENGTH, µm
0 1 2 3 4 5 6 7 8 9 10 11 12
RA
DIA
NC
E, 1
09 p
hoto
ns s
-1 c
m-2
ste
r-1 (
cm-1
)-1
1
10
100
1000
31%
23%
15%
10% (µ0 > 0.2)
120 K
110 K
100 K
130 K
θ0 = 67.5
θ0 = 0
• What we want to measure What happens when we try
Radiation Environment at Europa
CHARGE, fC
1 10
DIF
FE
RE
NT
IAL
HIT
RA
TE
, s-1
fC-1
0.1
1
10
100
1000
EUROPA G1ENHiLatInSb, DETECTOR NO.14
AVERAGE CHARGE
D&P Instruments TurboFT• Very rugged, only one moving (rotary) part• Data acquisition triggering• Spectral resolution is a function of the rotor
thickness and rotor material index of refraction• 4 quadrants for design purposes
FTIR Spectrometer• Multiplexing (Felgett) advantage, light at all wavelengths is collected
simultaneously.
• Higher light-gathering power than dispersive spectrometers (Jacquinot advantage). TurboFT (D&P Instruments) is approximately 50 × more sensitive in light gathering than the Galileo Near Infrared Mapping Spectrometer (NIMS).
• Built-in Radiation tolerance – Noise is spread and apodization
• AC-coupled to data acquisition
• Signal Processing can greatly increase performance in high radiation environment
Data Analysis for the TurboFT Spectrometer
• Peculiarities of the TurboFT – 4 quadrants
• Compute spectra by FFT
• Co-Add spectra for each quadrant
• Combine for a single spectrum
Data Rates and Resolution
• Acquire multiple scans– Between 10 and 360 scans per second – Multi-pixel versions
• Good spectral resolution 8 cm-1 (4096 interferogram data points), tighter if desired.
• At slow rotation speeds (10 scans/sec) the single-pixel data rate is approximately 1 MB/sec.
Integration Times and Data Redundancy
• Integration times estimated at approximately 60 seconds per physical location at Europa
• With 4 rotational positions, there are 150 (60*10/4) samples of each interferogram data point
• Time interval between successive interferogram points is approximately 8 us
Requirements for Signal-to-Noise are dependent on analysis
• Single wavelength analysis– Often what is taught at school– Deservedly bad reputation
• Integrated peaks– A modest improvement
• Spectral Methods– If there is structure in the target spectrum– Demonstrated SNR < 1 with high accuracy
• This talk is not about Spectral Processing Methods but about separable signal and noise in the measurement – a prelude to Spectral Processing
Statement of the Problem
• Each Interferogram is a sampling of the same signal so there are redundant samples
• Each sample contains signal and ambient noise measured together
• Because the data rate is fast compared to the radiation noise frequency near Jupiter, the “signal” mean can be recovered
Indexed Statistics
• Simple idea – Powerful in practice: Throw out what is inconsistent to find
what you are looking for
• Start with the median value of the distribution and compute mean value and standard deviation, throw out samples that are outside 2 sigma limits
• Repeat by re-sampling remaining data until convergence (usually
Signals and NoiseRange of Application and Limitations
• Signals must be separable:– Standard deviation of “signal” must be small
compared to magnitude of noise
• Examples:– Signal with Big noise– Signal with White noise– Signal with Small noise– Signal with Europa noise
Signal, Noise and MeasurementBig Noise
Normalized Histogram of Signal, Noise and Measureme ntSignal Mean at 10.0 - Idx Calculated Mean = 10.01 4
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0 5 10 15 20 25 30
Signal Intensity
Rel
ativ
e F
requ
ency
of O
ccur
renc
e Signal
Noise
Measurement Average = 15.09
Signal, Noise and Measurement Data
0
5
10
15
20
25
30
0 100 200 300 400 500 600
Sample Number
Sam
ple
Val
ue
Signal, Noise and MeasurementWhite Noise
Normalized Histogram of Signal, Noise and Measureme ntSignal Mean at 10.0 - Idx Calculated Mean = 10.04 3
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0 5 10 15 20 25 30
Signal Intensity
Rel
ativ
e F
requ
ency
of O
ccur
renc
e
Signal
Noise
Measurement Average = 15.62
Signal, Noise and Measurement Data
0
5
10
15
20
25
30
35
40
45
0 100 200 300 400 500 600
Sample Number
Sam
ple
Val
ue
Signal, Noise and MeasurementSmall Noise
Normalized Histogram of Signal, Noise and Measureme ntSignal Mean at 10.0 - Idx Calculated Mean = 10.14 5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0 5 10 15 20 25 30
Signal Intensity
Rel
ativ
e F
requ
ency
of O
ccur
renc
e
Signal
Noise
Measurement Average = 11.06
Signal, Noise and Measurement Data
-5
0
5
10
15
20
0 100 200 300 400 500 600
Sample Number
Sam
ple
Val
ue
Signal, Noise and MeasurementEuropa Noise (synthetic)
Signal, Noise and Measurement Data
0.0
0.5
1.0
1.5
2.0
2.5
0 100 200 300 400 500 600
Sample Number
Sam
ple
Val
ue
Normalized Histogram of Signal, Noise and Measureme nt Signal Mean at 0.50 - Idx Calculated Mean = 0.497
0
0.1
0.2
0.3
0.4
0.5
0.6
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
Signal Intensity
Rel
ativ
e F
requ
ency
of
Occ
urre
nce
Signal
Noise
Measurement - Mean = 0.79
Interferogram Jitter and Registration
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1950 1970 1990 2010 2030 2050 2070 2090
Interferogram Sequence Point
Signa
l Inten
sity / Volt
-1
0
1
0 500 1000 1500 2000 2500 3000 3500 4000
Inteferogram Data Point Sequence Number
Radiation Effects – 14 Interferograms
-1
0
1
2
3
4
5
6
7
8
0 500 1000 1500 2000 2500 3000 3500 4000
Inteferogram Data Point Sequence Number
Vol
tage
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
1900 1920 1940 1960 1980 2000 2020 2040 2060 2080 2100
Inteferogram Data Point Sequence Number
Vol
tage
-1
0
1
0 500 1000 1500 2000 2500 3000 3500 4000
Inteferogram Data Point Sequence Number
2 Sigma Limits
Registered
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1950 1970 1990 2010 2030 2050 2070 2090
Interferogram Sequence Point
Sig
nal I
nten
sity
/ V
olt
Unregistered
-2.5
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
1950 1970 1990 2010 2030 2050 2070 2090
Interferogram Sequence Point
Sig
nal I
nten
sity
/ V
olt
Application to Radiation in FTIR Data
0.0000
0.0002
0.0004
0.0006
0.0008
0.0010
0.0012
0.0014
0.0016
0.0018
0.0020
4 6 8 10 12 14 16 18 20
Wavelength - um
Sig
nal I
nten
sity
With Radiation
Without Radiation
Repaired
Another Example
0.0004
0.0005
0.0006
0.0007
0.0008
0.0009
0.0010
0.0011
0.0012
6.2 6.4 6.6 6.8 7.0 7.2 7.4 7.6 7.8 8.0
Wavelength - um
Sig
nal I
nten
sity
Radiation
New
Fixed
Removed fromIntegral
Integral from Wavelength = 6.5702 to 6.8196 um above the black trapezoid is: 0.0000129 for the original spectrum and 0.0000125 for the spectrum with Radiation that had been Fixed - 3.2% error in the feature bump
An Inverse Application to Curve Fitting of Noisy Data
• Fluorescence Background Removal in RamanSpectroscopy – where this numerical technique was originally developed
– Compute fluorescence signal, e.g., as Gaussian (3 variables)
– Compute Error and throw out big points with large error (Raman signal)
– Optimize fit of Gaussian to reduced data set
Removal of Raman Background Fluorescence
-1.0
0.0
1.0
2.0
3.0
4.0
5.0
6.0
0 1 2 3 4 5 6 7 8
Wavelength (arbitrary units)
Sig
nal I
nten
sity
Measured Signal
Computed FluorescenceBackground
Computed Raman Signal
Take-Aways
• FTIR has inherent tolerance for Radiation
• Instrument design can influence operation
• Data analysis can pull much information out of some types of noise
JPL Jovian References• http://opfm.jpl.nasa.gov/ • http://opfm.jpl.nasa.gov/europajupitersystemmissionejsm/ejsmpresentations• Boldt, J., et al., 2008. Assesment of Radiation Effects on Science and
Engineering Detectors for the JEO Mission Study. Jupiter Europa Orbiter Mission Study 2008: Final Report. JPL D-48256
• Carlson, R. W., 2010. Radiation Noise Effects at Jupiter: Comparison of In-situ and Laboratory Measurements RTD Final Report. Jet Propulsion Laboratory, Pasadena.
• Carlson, R. W., et al., 2009. Europa's Surface Composition. In: EUROPA (R. T. Pappalardo, et al., Eds.). Univ. Ariz. Press, Tucson, 283-327.
• Fieseler, P. D., et al., 2002. The radiation effects on Galileo spacecraft systems at Jupiter. IEEE Transactions on Nuclear Science. 49, 2739-2758.
• Hand, K. P., et al., 2009. Astrobiology and the Potential for Life on Europa. In: EUROPA (R. T. Pappalardo, et al., Eds.). Univ. Ariz. Press, Tucson, 589-630.
TurboFT References• Wadsworth, W., Dybwad, J. P., 1997. Ultra high speed chemical imaging spectrometer.
ElectroOptical Technology for Remote Chemical Detection and Identification, Vol. 3082. Soc. Photog. Instrum. Eng., pp. 148-154.
• Wadsworth, W., Dybwad, J. P., 1998. A very fast imaging FT spectrometer for on line process monitoring and control. Electro-Optic, Integrated Optic, and Electronic Technologies for Online Chemical Process Monitoring, Vol. 3537. Soc. Photog. Instrum. Eng., pp. 54-61
• Wadsworth, W. and Dybwad, J.P., 2001a, Airborne Testing of Small, Fast, Rugged Fourier Transform Spectrometer for Geologic Survey Use, Proceedings of Fifth International Airborne Remote Sensing Conference, San Francisco, CA, 17-20 September, 2001.
• Wadsworth, W., Dybwad, J. P., 2001b. Field testing of a small, fast, rugged Fourier transform spectrometer in the air and on the ground. ISSSR 2001, Quebec City, Canada.
• Wadsworth, W., Dybwad, J. P., 2001c. Rugged high speed rotary imaging Fourier transform spectrometer for industrial use. Vol. 4577. Soc. Photog. Instrum. Eng.
• Winthrop Wadsworth, "8x8 element mosaic imaging FT-IR for passive standoff detection“, 7th 2006 Standoff Detection Conference in Williamsburg, VA, 23-27 October 2006, Proc. SPIE 6302, 630202 (2006)
• "8X8 Element Mosaic Imaging FT-IR for Passive Standoff Detection“, SPIE Optics & Photonics Conference in San Diego, 13-17 August, 2006
• Hewson, R., et al, Hyperspectral Thermal Infrared Line Profiling for Mapping Surface Mineralogy, Proceedings of Fifth International Airborne Remote Sensing Conference, San Francisco, CA, 17-20 September, 2001.
• http://www.dpinstruments.com/publications.php