ASPECT Technical Description
The ASPECT sensor suite (Figure 1) is mounted
in a fixed wing aircraft and uses the principles of remote hazard detection to
image, map, identify, and quantify chemical vapors and deposited
radioisotopes. Chemical plume measurements
are made at a rate of about two square miles per minute. The system normally operates at an altitude
of 2800 feet above ground level (AGL) and results in a high IR spatial
resolution of 0.3 meters. A simplified
system diagram is provided in Figure 2.
The radiological data is collected between 300 and 500 feet AGL with a
collection time of once per second and a field of view about 600-1000 feet. Situational awareness is provided by using
both high-resolution aerial digital photography and digital video that are
concurrently collected with chemical and radiological data and forms the basis
for a geographical information system data cube with several layered data
products (Figure 3). Efficient mission
execution requires that data is processed on-board the aircraft for
transmission or hand-off to the first responder. To facilitate data transmission while in
flight status, the aircraft is equipped with a broadband high-speed satellite
data communications system. With a
combination of onboard data processing and the satellite communication system,
selected airborne situational data sets are ready for dissemination to the
incident command team in less than 5 minutes after collection.
FIGURE 1 – Sensor Suite
NOTE:
Images of the neutron detector and LaBr detectors are not shown. The aircraft is equipped with three RSX-4
units, only one is shown in this figure.
FIGURE 2 -
Simplified ASPECT System Diagram
FIGURE 3 – Data Products
1.
Airframe
The ASPECT sensor suite is
operated from a single engine Cessna 208B Cargo Master aircraft (Figure
4). The aircraft and crew are certified
for full instrument flight rules (IFR) flight operations. This aircraft is equipped with one 20 X 30
inch belly hole with a retractable bay door.
All sensor systems are mounted on vibration isolated base plates
positioned over the belly hole. The
aircraft can operate from any airport having a 3000 ft runway and can stay
aloft for 5 hours. Technical
specifications for the program airframe are contained in Table 1.
FIGURE 4 –
ASPECT Aircraft, N9738B
TABLE 1 –
N9738B Technical Specifications
Tail: |
N7938B |
System: |
Single
Engine Turbo-Propeller Driven Aircraft |
Make: |
Cessna 208B
Cargo Master, Part 91 Certification |
Power
Plants: |
Pratt and
Whitney PT-6A-114 turbine driven three blade propeller, 600 shaft HP |
Empty
Weight: |
4458 lbs. |
Useful
Weight: |
4604 lbs. |
Maximum
Take-off Weight: |
9062 lbs. |
Typical
Cruise Speed: |
160 Kts |
Typical
Flight Duration: |
6 Hours
(65% Power) Plus 45 Minute Reserve |
Service and
Ceiling: |
Low
Altitude Waiver, 25000 Ft (MSL) max altitude |
Cabin: |
Un-pressurized,
Crew Oxygen |
Portals: |
One, 20 X
30 inch STC sensor hole with Remote
Door |
Avionics |
GPS IFR
Package GPS
navigation system for programmed flight lines Terrain/Obstacle
Avoidance Equipped Radar
Altimeter Equipped Dual
weather Radar, Live Weather Feed Dual VOR
Equipped, Dual Comms Equipped Dual
Transponders |
Electrical
Buss: |
Primary, 28
vdc @ 200 amps full load Secondary,
28 vdc @ 90 amps full load |
Data
Communication: |
Phased
Array Satellite System, 40-60 KB/sec Data/Telephone combined |
Readiness
Status: |
24/7/365 |
2.
Chemical Detection Capabilities
The principle of remote
detection, identification, and quantification of a chemical vapor species is
accomplished using passive infrared spectroscopy. Most vapor compounds have unique absorption
spectral bands at specific frequencies in the infrared spectral region. An asymmetric stretching between two atoms in
a molecule results in a fundamental frequency of vibration. Passive infrared measurements of a vapor
species are possible due to small thermal radiance differences between the
temperature of the chemical plume and a particular infrared scene background
(Figure 5). Both the cloud and the
atmosphere contribute to the total emitted radiance measured by an infrared
sensor. Careful monitoring of the change
in total infrared radiance levels leads to concentration estimations for a
particular vapor species. Concentration
times path length estimations are obtained based on the molar absorptivity for
each vapor species.
FIGURE 5 -
Principles of Remote Infrared Detection
A.
RS800IRLS Infrared Imager Sensor
The ASPECT Program uses a
modified Raytheon RS800 infrared line scanner to generate wide area chemical
imagery (Figure 6). This system
incorporates a unique detector assembly consisting of 16 cryogenically cooled
optical band pass filters affixed to a focal plan array. Scanning is accomplished with an integrated
rotating prism controlled by a feedback motor scan controller. Each rotation of the prism sweeps an angular
field of view of 60 degrees resulting in 1500 data points. When the scan rate is coupled to the normal
110-knot ground speed of the aircraft, a scan swath of 0.5 meters is
collected. This collection geometry
outputs a square data pixel 0.5 X 0.5 meters square. Radiometric calibration is performed during
each prism rotation by viewing two reference blackbodies mounted on either side
of the unit. Image registration is
accomplished during post processing by incorporating pitch and roll data
collected from an integrated gyroscope mounted on the scanner body. An integrated GPS receiver is used in the
processing step providing geo-registration of each pixel in the finished image
product. Detailed specifications of the
RS800IRLS are contained in Table 2.
FIGURE 6 -
RS800IRLS Line Scanner
TABLE 2 -
RS800IRLS Technical Specifications
System: |
TI
Systems/Raytheon RS800 MSIRLS |
Detector: |
Cryogenically
cooled focal plane array with integrated cold optical filters |
Spectral
Coverage: |
3 – 5
micrometer (mid-wave) and 8 – 12 micrometer (long-wave) |
Number of
Spectral Channels: |
16 total, 8
mid-wave and 8 long-wave |
Spectral
Resolution: |
5 to 20
wave numbers, channel dependent |
Spatial
Resolution: |
Better than
1.0 mill radian |
Scan Rate: |
60 Hz |
Radiometric
Calibration: |
Two
flanking blackbody units |
Field of
View (FOV): |
60 Degrees |
Thermal Resolution |
0.05 Degree
C. |
Linear
Range |
0 to 200
Degree C |
Pixel
Resolution (IFOV): |
0.5 meters
@ 850 meter collection altitude (AGL) |
Cross Field
Scan Coverage: |
980 meters
@ 850 meter collection altitude (AGL) |
Attitude
Stabilization: |
25 Hz pitch
and roll providing stabilized video |
Power: |
28 vdc @ 10
amps full load |
Weight: |
27 Kg (60
lbs.) |
Spin-up
Time: |
Less than
12 minutes (including cyro-system) |
Standard
Outputs: |
2 Channels
of stabilized RS-170 video, 16 channels of digitized (16 bit) spectral data,
1 channel of GPS (2 Hz) |
Data
Processing |
1 step full
radiometric image generation using an onboard algorithm. Approximately 1 minute processing time. |
B.
LS1600 Infrared Imager Sensor
A replacement infrared
imaging system is scheduled to be incorporated into the ASPECT system in December
2016 (figure 7). This system contains 16
regularly spaced long-wave channels ranging in spectral resolution of 15 to 30
wave numbers. By carefully selecting the
placement and bandwith of these channels, the program is constructing a low
resolution spectrometer permitted pattern recognition techniques to be applied
to the resulting imagery. These system
enhancements (over the RS800) will provide a factor of 2 to 5 sensitivity
improvement translating to chemical image detection at much lower
concentrations. Design and interim
construction specifications are provided in table 3.
Figure 7 --
LS1600 Infrared Line Scanner Imaging System
TABLE 3 –
LS1600 Infrared Imager Technical Specifications
System: |
LR Tech
LS1600 Line Scanner |
Detector: |
Cryogenically
cooled focal plane array with integrated cold optical filters |
Spectral
Coverage: |
8 – 13.5 micrometer (long-wave) |
Number of
Spectral Channels: |
16 total,
all long-wave |
Spectral
Resolution: |
15 to 30
wave numbers, regularly spaced |
Spatial
Resolution: |
Better than
1.0 mill radian |
Scan Rate: |
60 Hz |
Radiometric
Calibration: |
Two
flanking blackbody units |
Field of
View (FOV): |
60 Degrees |
Thermal
Resolution |
0.05 Degree
C. |
Linear
Range |
0 to 200
Degree C |
Pixel
Resolution (IFOV): |
0.5 meters
@ 850 meter collection altitude (AGL) |
Cross Field
Scan Coverage: |
980 meters
@ 850 meter collection altitude (AGL) |
Attitude
Stabilization: |
1600 Hz pitch
and roll providing stabilized video |
Power: |
28 vdc @ 10
amps full load |
Weight: |
27 Kg (60
lbs.) |
Spin-up
Time: |
Less than
12 minutes (including cyro-system) |
Standard
Outputs: |
1 Channel
of stabilized RS-170 video, 16 channels of digitized (18 bit) spectral data,
with associated attitude and GPS data (5 Hz) |
Data
Processing |
1 step full
radiometric image generation using an onboard algorithm. Approximately 1 minute processing time. |
C.
IRLS - Chemical Image Processing
1.
Chemical Processing
Processing of chemical data
is divided into two broad categories including image processing and spectral
processing. Infrared chemical signatures
present a challenge in data processing due to the small signal to noise ratio
(SNR) of the chemical vapor between the sensor and the surface. It is not uncommon to have a SNR of less than
four in a typical vapor cloud. In order
to image such a weak signal, the collection system and detector must be
optimized for high collection efficiency and a small instantaneous field of
view. The ASPECT RS800IRLS meets both of
these requirements by using an F/1 high-speed optical train coupled to a
16-channel cold optical filter focal plane array. This configuration provides very high signal
throughput while maintaining a 1.0 mill radian spatial resolution. The use of cold optical band pass filters
directly mounted on the face of the focal plane array eliminates a large
portion of the self-radiance (noise) while minimizing the attenuation of wanted
signal content. Raw data is fully
radiometrically calibrated using a set of flanking blackbodies providing
radiance-adjusted imagery. Jitter
removal and band registration are accomplished using an automated algorithm
using an integrated 2 dimensional gyro and GPS feed. Final data is generated using an automated
geo-registration algorithm in either a geo-Tiff or a geo-Jpeg format. Processing can be accomplished while in
flight and requires about 30 seconds to 1 minute per image depending on
size. The completed imagery is compressed
and made available to ground user through the aircraft satellite link.
The vapor cloud, shown in
Figure 8, is captured in an infrared image collected by the RS800SIRLS
multispectral sensor at an industrial site in the Midwest. The detection limit of the vapor
concentration (shown in red) was determined to be less than 20 ppm/m while the
center cloud concentration was greater than 250 ppm/m. This image has been cropped with only 1/3 of
the actual sensor field-of-view being displayed. The original image width of
the image was approximately 1200 meters wide.
For the ASPECT application, the RS800IRLS system provides a qualitative
indication of the presence or
absence of a particular
chemical species. The detection limit
provided by the sensors is applicable to both chemical emergency response and
crisis mitigation following a terrorism event.
2.
Thermal Processing
Figure 9 shows a long wave
thermal image of an underground mine fire in Kansas City, MO. In this application, the calibrated
radiometric data from the RS800IRLS is processed to show direct surface
temperature. In the standard
configuration, the imager provides an automated linear temperature measure up
to about 200 oC at a
resolution of 0.05 oC. Higher temperature measurements in the order
of 500 oC are possible
by changing the input stop and detector of the instrument. As part of the thermal processing method,
solutions to the scene emissivity are solved and factored into the temperature
product.
FIGURE 9 - Thermal Image of an
Underground Mine Fire
3.
Oil
Detection Processing
The ASPECT sensor system was
extensively used to detect oil on water during the Deep Water Horizon Oil
spill. The application of long wave IR
to detect oil has been known for a number of years and has advantages over
visible detection of oil including the ability to image oil under strong sun
angle (glint) and the ability to image oil at night. The primary problem of using long wave IR in
oil detection is centered on an accurate measure of the surface temperature with
a pixel size that is small enough to provide adequate scene dynamic range. Figure 10 shows an image that was processed
from oil on water in the open ocean. In
this setting, the surface temperature of the ocean is very constant. The image was formed by detecting the
difference in emissivity between 0.96 for water and 0.93 for oil. Weathered oil has an emissivity of 0.98 or
0.99 and was usually at a different surface temperature. This image was processed using the small
emissivity differential of oil and water as the primary discriminate, coupled with
weather oil having a different surface temperature.
By using a processing
technique that examined data value boundaries, the amount of oil on the water
surface was generated. Figure 11 shows
an image of oil on water in a near shore environment. For this environment, the thermal gradient of
the water in the shallows is much greater than the emissivity contrast of the
oil and water. A multi-dimensional
pattern recognition algorithm utilizing a pre-processing procedure known as
alpha residuals to eliminate the temperature difference was used to
discriminate oil and water. This
technique permitted detection and characterization of oil ranging from sheen to
thick oil.
FIGURE 10 - Open Water
Detection and Discrimination of Oil Using IR
FIGURE 11 - Detection and
Discrimination of Oil using Pattern Recognition
D.
MR254AB Spectrometer
Chemical vapor detection and
quantification is accomplished using a modified Bomen MR254AB spectrometer
(Figure 12). This custom designed
spectrometer utilizes a double wishbone pendulum interferometer providing both
high signal throughput and vibrational noise immunity. Two cryogenically cooled detectors provide
both mid and long wave operation.
Spectral resolution is selectable and ranges from 1 to 128 wave number
with 16-wave number normally used for automated compound detection. When operated at 16 wave number resolution,
the unit scans at 70 Hz providing a spatial sampling interval every 0.75 meters
along the ground track of the aircraft.
The program uses an automated compound detection algorithm based on
digital filtering and pattern recognition.
Geo-registration of each Fourier Transform Spectrometer (FTS) scan is
accomplished using a concurrent GPS input from the ASPECT main GPS receiver. Technical specifications of the FTS system
are contained in Table 4.
FIGURE 12 -
MR254AB FTS Spectrometer
TABLE 4 -
FTS Technical Specifications
System: |
Modified
Bomem MR-200 Series (MR-254AB) |
Detectors: |
Cryogenically
cooled Single Pixel Design |
Spectral
Coverage: |
InSb
Detector for 3 – 5 micrometer (mid-wave) MCT Detector for 8 – 12 micrometer
(long-wave) |
Noise
Figure |
Mid-wave
6x10-9 W/cm2-srcm-1, Long-wave 1.8x10-8
W/cm2-srcm-1 |
Spectral
Resolution: |
1 to 128
wave numbers, User selectable |
Spatial
Resolution: |
5
milli-radian (0.2o) thru 25.4 cm (10 inch) Primary Telescope, 0.75 meter interval at 110 kts collection
velocity |
Scan Rate: |
70 Hz @ 16 wave number resolution |
Field of
View (FOV): |
3 meters @
850 meter collection altitude (AGL) |
Radiometric
Reference |
Integrated
cold source (77o K) |
Targeting |
Calibrated
bore camera (Nadir) |
Power: |
28 vdc @ 8
amps full load |
Weight: |
40 Kg (90
lbs.) |
Spin-up
Time: |
Less than 4
minutes (including cyro-system) |
Standard
Outputs: |
2 Channels
of Grams format spectral data (16 bit), 1 channel of RS-170 video. |
Data Processing: |
1 Step
pattern recognition compound detection using onboard algorithms, approx. 30 –
60 seconds processing time after data collection. |
E.
VSR Spectrometer
The MR254AB spectrometer is
scheduled to be replaced in January 2017 with a two channel Versatile
SpectroRadiometer (VSR) (figure 13). The
VSR consist of a double wishbone interferometer with integrated calibration
blackbodies. The program is currently
conducting noise improvements to the
system by out boarding the cyro-pumps and upgrading the long wave detector to
include both a top hat and cold filter configuration. The goal of these activities is to improve
the noise figure of the instrument to a 2 to 5fold improvement over the
MR254AB Design and interim construction specifications
are provided in table 5.
Figure 13 –
VSR Spectrometer
TABLE 5 -
VSR Technical Specifications
System: |
VSR Double
Wishbone FTS System |
Detectors: |
Cryogenically
cooled Single Pixel Design |
Spectral
Coverage: |
InSb
Detector for 3 – 5 micrometer (mid-wave) MCT Detector for 8 – 12 micrometer
(long-wave) |
Noise
Figure |
Mid-wave
4.5x10-10 W/cm2-srcm-1, Long-wave
2.3x10-9 W/cm2-srcm-1 |
Spectral
Resolution: |
1 to 128
wave numbers, User selectable |
Spatial
Resolution: |
5
milli-radian (0.2o) thru 25.4 cm (10 inch) Primary Telescope, 0.75 meter interval at 110 kts collection
velocity |
Scan Rate: |
70 - 90 Hz @ 16 wave number resolution, rate
selectable |
Field of
View (FOV): |
3 meters @
850 meter collection altitude (AGL) |
Radiometric
Reference |
Integrated
blackbody sources (303o K and 328o K ), calibrated
after each measurement |
Targeting |
Calibrated
bore camera (Nadir) |
Power: |
28 vdc @ 8
amps full load |
Weight: |
40 Kg (90
lbs.) |
Spin-up
Time: |
Less than 4
minutes (including cyro-system) |
Standard Outputs: |
2 Channels
of Grams format spectral data (16 bit) |
Data
Processing: |
1 Step
pattern recognition compound detection using onboard algorithms, approx. 30 –
60 seconds processing time after data collection. |
F.
Spectral Processing
Spectral data processing
(signal processing) from the ASPECT MR-254AB spectrometer is processed using
background suppression, pattern recognition algorithm. Processing spectral data from a moving
airborne platform requires unique methods to balance weak signal detection
sensitivity, false alarm minimization, and processing speed. The background suppression, pattern
recognition methods associated with the ASPECT Program have been documented in
over 100 open literature publications.
One of the principal
weaknesses of airborne FTS data is the ability to reference each collected
spectra to a suitable background for subsequent spectra subtraction. While methods have been devised to accomplish
this procedure, typical airborne spectra show changes between successive scan
of several orders of magnitude due to changing radiometric scene
conditions. These scan-to-scan changes
render traditional background subtraction methods unusable for weak signal detection. The background suppression method used by
ASPECT circumvents this problem by using a digital filtering process to remove
the background component from the raw interferometer data. This approach is analogous to using the
tuning section of a radio receiver to preselect the portion of the signal for
subsequent processing. The resulting
filtered intermediate data maintains the weak signal components necessary for
subsequent analysis.
An additional weakness of
traditional FTS processing involves the need to provide a resolution high enough
to permit compounds exhibiting narrow spectra features to be matched with
published library spectra. This method
is initiated using the Fast Fourier Transform (FFT). While the FFT algorithm is very robust
mathematically, certain data collection requirements must be met to permit the
transform to be valid. In order to
provide high spectral resolution spectra, the length of the interferogram must
be matched to the desired resolution for the transform to work properly. This requirement forces a long collection
period for each interferogram and since the aircraft is moving, it is probable
that the radiometric scene being viewed by the spectrometer will change during
the collection of the interferogram. The
changing scene causes the FFT to generate spectral artifacts in the resulting
spectral information. These artifacts
are phantom signals that confuse and complicate subsequent compound
identification.
The standard matched filter
compound discrimination method likewise exhibits weak signal performance and
often generates false alarms due to common atmospheric interference. ASPECT solves these problems by using a
combination of digital band pass filtering followed by a multi-dimensional
pattern recognition algorithm. The
digital filters and pattern recognition coefficients are developed using a
combination of laboratory, field, and library data and folded into a training
set that is run against unknown data.
Digital filters can be readily constructed which take into account both
spectrometer line shapes and adjacent interferents, greatly improving the weak
signal system gain. The pattern
recognition algorithm processes the filter output in a multi-space fashion and
enhances the selectivity of the detection.
These methods are very similar to a superheterodyne receiver that uses
band pass adjustable intermediate filters followed by a DSP
detector/discriminator such as in a modern radar system. Since the methods use relatively simple
computational operations, signal processing can be accomplished in a few
seconds. Finally, as data is processed,
the position of the detection is referenced to onboard GPS data providing a GIS
ready data output. Table 6 lists the
compounds (and a 10 meter equivalent path length detection limit) that are
currently installed in the airborne library using the digital filtering/pattern
recognition method.
A unique feature of the
ASPECT System includes the ability to process spectral data automatically in
the aircraft with a full reach back link to the program QA/QC program. As data is generated in the aircraft using the
pattern recognition software, a support data package is extracted by the reach
back team and independently reviewed as a confirmation to data generated by the
aircraft.
Figure 14 shows airborne
absorbance spectra of ammonia vapor collected from an ammonium nitrate fire
using the MR-254AB spectrometer. This
spectrum was generated by carefully selecting a suitable background spectra and
conducting a traditional background subtraction, a time consuming operation. Figure 15 shows the same data processed using
the automated background suppression/pattern recognition method. Ammonia detection is clearly
demonstrated. Figure 16 shows how these
detections are referenced to a real-world geographical map. Individual detection locations corresponding
to FTS scans are mated with latitude and longitude coordinate values.
TABLE 6 - Chemicals
Included in the ASPECT Auto-Processing Library
Acetic Acid
(2.0) |
Cumene
(23.1) |
Isoprene
(6.5) |
Propylene
(3.7) |
Acetone
(5.6) |
Diborane
(5.0) |
Isopropanol
(8.5) |
Propylene
Oxide (6.8) |
Acrolein
(8.8) |
1,1-Dichloroethene
(3.7) |
Isopropyl
Acetate (0.7) |
Silicon
Tetrafluoride (0.2) |
Acrylonitrile
(12.5) |
Dichloromethane
(6.0) |
MAPP (3.7) |
Sulfur
Dioxide (15) |
Acrylic
Acid (3.3) |
Dichlorodifluoromethane
(0.7) |
Methyl
Acetate (1.0) |
Sulfur
Hexafluoride (0.07) |
Allyl
Alcohol (5.3) |
1,1-Difluoroethane
(0.8) |
Methyl
Ethyl Ketone (7.5) |
Sulfur
Mustard (6.0) |
Ammonia
(2.0) |
Difluoromethane
(0.8) |
Methanol
(5.4) |
Nitrogen
Mustard (2.5) |
Arsine (18.7) |
Ethanol
(6.3) |
Methylbromide
(60) |
Phosgene
(0.5) |
Bis-Chloroethyl
Ether (1.7) |
Ethyl
Acetate (0.8) |
Methylene
Chloride (1.1) |
Phosphine
(8.3) |
Boron
Tribromide (0.2) |
Ethyl
Formate (1.0) |
Methyl
Methacrylate (3.0) |
Tetrachloroethylene
(10) |
Boron Triflouride
(5.6) |
Ethylene
(5.0) |
MTEB (3.8) |
1,1,1-Trichloroethane
(1.9) |
1,3-Butadiene
(5.0) |
Formic Acid
(5.0) |
Naphthalene
(3.8) |
Trichloroethylene
(2.7) |
1-Butene
(12.0) |
Freon 134a
(0.8) |
n-Butyl
Acetate (3.8) |
Trichloromethane
(0.7) |
2-Butene
(18.8) |
GA (Tabun)
(0.7) |
n-Butyl
Alcohol (7.9) |
Triethylamine
(6.2) |
Carbon
Tetrachloride (0.2) |
GB (Sarin)
(0.5) |
Nitric Acid
(5.0) |
Triethylphosphate
(0.3) |
Carbonyl
Chloride (0.8) |
Germane
(1.5) |
Nitrogen
Trifluoride (0.7) |
Trimethylamine
(9.3) |
Carbon
Tetraflouride (0.1) |
Hexafluoroacetone
(0.4) |
Phosphorus
Oxychloride (2.0) |
Trimethyl
Phosphite (0.4) |
Chlorodifluoromethane
(0.6) |
Isobutylene
(15) |
Propyl
Acetate (0.7) |
Vinyl
Acetate (0.6) |
FIGURE 14 - Ammonia
Spectra
FIGURE 15 -
Ammonia Detected with Pattern Recognition
FIGURE 16 - Locations of
Ammonia Detection
Quantitative compound
specific information is also generated using the MR-254AB spectrometer. This application uses a multi-dimensional
model generated using radiometric, thermal, and concentration calibrated
laboratory data for each compound in the airborne library. As with compound detection methods, multiple
publications have documented the feasibility of using this approach to remotely
quantify chemical vapors. The first open
literature scientific peer reviewed paper was completed using the ASPECT
method.
Figure 17 shows the estimated
concentration of methanol measured by ground sensors and compared with the
remotely collected FT-IR data. The data
shows a standard error of prediction of 18 ppm-m for a range of concentration
between 20 to 400 ppm-m. This range of
concentrations is consistent with both hazardous vapor releases and terrorist
concerns.
Figure 17 - Quantitative Methanol Results
3.
Radiological Detection Capabilities
A. RSX4 Sodium
NaI Gamma Ray Spectrometers
Airborne radiological
measurements are conducted using three fully integrated multi-crystal sodium
iodide (NaI) RSX4 gamma ray spectrometers (Figure 19). Each RSX4 spectrometer contains four
4”x2”x16” doped NaI crystals each having an independent photomultiplier/
spectrometer assembly. One RSX unit is
configured with an additional upward NaI crystal utilized to provide real-time
cosmic ray correction. Count and
energy data from each crystal and pack is combined using a self-calibrating
signal processor to generate a virtual detector output. All spectrometer “packs” are further combined
using a signal console controlled by the on-board computer in the
aircraft. Due to the advanced signal
processing techniques unique to the RSX4 units, very high total count rates
(approximately 1 million counts per second) can be discriminated and processed. Specifications of the RSX4 spectrometers can
be found in table 8.
B.
Lanthanum Bromide Gamma Ray
Spectrometers
High resolution gamma ray
detection is provided using a set of three 3” x 3” lanthanum bromide (LaBr)
crystals all ganged into a single virtual detector. This configuration of detector has a much
lower efficiency but has a much higher degree of spectral resolution than the
NaI spectrometers. Accordingly, the LaBr
crystal are most often used in complex isotope environments having a higher
intensity such as in a fall-out field resulting from a nuclear accident.
Neutron detection is provided
using a set of two, four bundle straw neutron detectors. These systems do not use expensive and rare
He3 gas and are very rugged. Both
detectors are ganged into one virtual total count sensor. Radiological spectral data, GPS position,
and radar altitude are collected at a one-second interval at all times during a
survey. In order to provide optimal
collection geometry, flight line data is loaded into the aircraft flight
computer prior to conducting the survey.
Typical airborne surveys are flown at 300 to 500 feet AGL.
FIGURE 19 -
RSX4 Gamma Ray Spectrometer
C.
Neutron Detection
Proper spectrometer operation
and data quality assurance is maintained using both internal and external
calibration algorithms. A self-contained
internal calibration algorithm acts as a watchdog and continuously monitors the
spectrometer systems for proper system operation and data output. If any errors are encountered with a specific
crystal and/or spectrometer pack during the collection process an error message
is generated and the data associated with that crystal are removed from further
analyses. External calibration
procedures are routinely executed and consist of both designed data collection
over characterized areas and pad calibrations over known quantities of
radiological doped concrete. Technical
specifications for the RsX4 gamma ray spectrometers are contained in Table 8.
TABLE 8 -
RX4 Technical Specifications
System: |
RSI RSX4
Gamma Ray Spectrometer |
Detector:
Gamma Ray |
4 Doped NaI
Detectors per pack, 3 Packs, 2x4x16 inch crystals, |
Detector:
Cosmic Ray |
One 2x4x16
upward crystal integrated into one of the RSX units. |
Total NaI Detector
Volume: |
25 liters |
Energy
Coverage: |
0 – 3000
KeV |
Number of
Channels: |
1024 |
Energy
Resolution: |
Approx. 3
KeV per Channel |
Scan Rate: |
1 Hz |
Internal
Calibration: |
Automatic
based on Natural K, U, and T |
Field of
View (FOV): |
45 Degrees |
Cross Field
Scan Coverage: |
300 meters
@ 300 meter collection altitude (AGL) |
Altitude
Determination: |
2.4 GHz
Radar Altimeter, 10 Meter DEM Database |
Power: |
28 vdc @ 4
amps full load (3 Packs) |
Weight: |
136 Kg (300
lbs.) |
Spin-up
Time: |
Less than 5
minutes |
Standard
Outputs: |
1024 Gamma
Ray Spectra, GPS (2 Hz) |
Data
Processing |
1 Step Full
Processing of Total Count, Sigma, and Exposure Rate, Approximately 1 minute Processing Time
After Data Collection. |
D.
Radiological Data Processing
All radiological data is
processed automatically using airborne algorithms. Normally, a specifically designed survey
flight path is flow by the aircraft and once complete, a suite of radiological
products is generated from the collected data.
Since radiological sources are universally present from the earth and
from cosmic sources, all radiological data must be corrected to establish a
baseline measurement. Cosmic estimates
are established by flying the aircraft 3000 feet AGL while collecting gamma
spectral data. At altitudes of 3000 feet
and greater all radiological inputs are either from the cosmic sources or the
aircraft (which is a constant).
Quantified cosmic contributions are stripped out from all subsequent
data. Depending on the length of the radiological
survey, cosmic backgrounds may be collected at the beginning and end of the
survey. In a fashion similar to the
cosmic correction, the natural radiological background for the survey area is
also established. This process normally
calls for collecting a limited amount of data (a test line) at the survey
altitude (300 – 500 ft. AGL) in an area of similar geology/land use but outside
of the region of survey interest. By
subtracting the test line data from the survey data, a corrected radiation map
for the survey area is generated.
Several data products are
generated automatically by the system including total counts, a sigma map, and
an exposure map. The total count product
is generated by mapping the corrected total count data (approximately 30 – 3000
KeV) from the spectrometers using the integrated GPS data as the geographic
datum. Exposure Rate mps are normally
contoured at regular intervals in micro-Roentgens (µR). Figure 20 illustrates a typical survey total
count plot.
FIGURE 20 -
Total Count Plot
A second radiological product
includes an array of isotope specific sigma plots or maps. These plots are very useful to the first
responder since they help highlight specific data points that may require
detailed ground investigation. This
procedure consists of a two-step method with the first being a windowing for
selected isotope energies followed by a statistical treatment of the data. Isotope specific data is generated by
windowing the gamma spectrum at energy levels corresponding to the isotopes of
interest. As part of this analysis,
higher energy contributions from uranium and thorium are removed using a
stripping coefficient. A statistical
average and standard deviation is next computed for the entire survey area
using the isotope windowed data. Since
the standard deviation provides a measure of the variance of the data set, data
values showing several standard deviations (sigma) indicate that these values
are statistically different from the majority of the population. ASPECT uses a graded scale in which 0 to 4
sigma are considered normal and greater than 4 sigma highlights data very
different from the population. Greater
than 6 sigma indicates that the data is extremely different and warrants
additional investigation. By using
different isotope windows, a number of sigma maps can be generated for a given
survey (Figure 21).
The final set of products
generated by the gamma ray spectrometers consist of an exposure plot or
map. This procedure consist of
extrapolating the measured total count data collected at the flight altitude
down to the total count that would be measured 1 meter above the surface. This method utilizes a weighting algorithm
that provides more focus on the high energy counts since these represent the
most energetic and penetrating gamma rays.
The extrapolation process is accomplished using the calibration
coefficients developed as part of the exterior calibration process. The resulting data is plotted in µR/hr and
provides the first responder with a health-based estimate of radiological
dosage at the ground surface (Figure 22).
FIGURE 21 -
Sigma Plot
FIGURE 22 -
Exposure Plot
4.
Aerial Camera Systems
A.
Nikon D2X
ASPECT utilizes a still
digital Nikon DX2 camera to collect and provide visible aerial imagery as part
of the core data product package (Figure 23). The DX2 consists of a 12.4 mega
pixel CMOS camera supporting a 3:5 aspect ratio frame. The system uses a 28 mm
wide-angle lens and is slaved to the primary IR sensors and provides concurrent
image collection when the other sensors are triggered for operation. All
imagery is geo-rectified using both aircraft attitude correction (pitch, yaw,
and roll) and GPS positional information. Imagery can be processed while the
aircraft is in flight status or approximately 600 frames per hour can be
automatically batch processed once the data is downloaded from the aircraft.
Technical specification for the DX2 camera is provided in Table 8.
Figure 23: ASPECT Camera Suite
TABLE 8 -
DX2
Aerial
Digital Camera Technical Specifications
System: |
Nikon DX2
Camera Body |
Detectors: |
12.4-megapixel
digital CMOS sensor |
Aspect
Ratio: |
3:5 |
Lens: |
28 mm
Digital Compatible |
Field of
View (FOV): |
824 meters
Cross flight and 548 meters Direction of Flight @ 850 meter collection
altitude (AGL) |
Pixel
Resolution (IFOV): |
19.2 cm @
850 meter collection altitude (AGL) |
Frame
Timing and Collection Rate: |
Operator
Selectable, 3 to 8 seconds, Approximately 600 frames per hour |
Trigger
Control: |
Automatic,
Manual, and Slave |
Power: |
12 vdc @ 1
amp full load |
Spin-up
Time: |
Less than 2
minutes from System Start |
Standard
Outputs: |
JPEG, Tiff |
Data
Processing: |
Full
INS/GPS Geospatial Rectification |
B.
Imperx MSIC
The Nikon DX2 is scheduled to
be replaced in December 2015 with an Imperx mapping camera (figure 24). The new mapping camera will provide a similar
aspect ratio and aerial coverage at a much higher resolution. Due to the size of the CCD sensor, little
edge distortion will be present in the frames.
Technical specification for the mapping camera is shown in table 9.
Figure 24 –
Imperx Mapping Camera
TABLE 9 -
Imperx Aerial Digital Camera Technical Specifications
System: |
Imperx
B6640 body |
Detectors: |
29-megapixel
digital CCD sensor (KAI-29050) |
Aspect
Ratio: |
4:5 |
Lens: |
24 mm
Digital Compatible |
Field of
View (FOV): |
824 meters
Cross flight and 548 meters Direction of Flight @ 850 meter collection
altitude (AGL) |
Pixel
Resolution (IFOV): |
92 cm @ 850
meter collection altitude (AGL) |
Frame
Timing and Collection Rate: |
Operator
Selectable, 1 to 15 seconds, Approximately 600 frames per hour for normal
mission |
Trigger
Control: |
Automatic,
Manual, and Slave |
Power: |
12 vdc @ 1
amp full load |
Spin-up
Time: |
Less than 2
minutes from System Start |
Standard
Outputs: |
JPEG,
TrueSense |
Data
Processing: |
Full
INS/GPS Geospatial Rectification |
C.
Canon EOS Oblique Camera
In order to provide
situational information from the perspective of the flight crew, ASPECT also
supports an oblique camera system that is operated from the right side of the
aircraft. This camera consists of a Canon EOS Rebel digital SLR camera body
with a 30 – 120 mm variable zoom lens (Figure 25). Frames are collected at an
approximate the 2 o’clock position relative to the aircraft with the target
approximately 1000 meters from the aircraft. Figure 7 provides examples of an
aerial (downward view) photo and an oblique (side view) photos. The aerial
photos are taken at an altitude of about 2,800 feet above the ground (AGL) and
the oblique photos are taken at lower altitudes ranging from 500 feet to about
1,200 feet AGL. Table 3 provides technical specification of the oblique camera
system.
Figure 25
-- Canon EOS Rebel Digital SLR Camera
TABLE 10 - Canon EOS Aerial Oblique
Digital Camera Technical Specifications
System: |
Canon
EOS Camera Body |
Detectors: |
6.3-megapixel
digital CMOS sensor |
Aspect
Ratio: |
3:5 |
Lens: |
30-120
mm zoom, Digital Compatible |
Trigger
Control: |
Manual |
Power: |
Internal
Battery |
Standard
Outputs: |
JPEG,
Tiff |
Data
Processing: |
Spatial
Geo-reference |
D.
Visible Imagery Data Processing
Visible imagery collected
with the ASPECT System is ultimately processed into a geo-registered jpeg or
tiff format image. Image processing is
composed of two primary steps including image enhancement and geo-registration. Both of these processing steps can be
processed while the aircraft is in flight status but typically, imagery is
processed once the aircraft lands due to the large quantity of data involved
with aerial photography. A standard
flight mission often generates 600 aerial images.
The ASPECT aerial camera
consists of a still frame 3x4 ratio digital camera. A wide field of view lens is utilized to
match the ground width coverage of the line scanner system. Due to the speed of the aircraft and the fact
that ASPECT may fly in low light conditions, the camera uses a fixed focus and
shutter speed configuration. Raw imagery
is subsequently processed to balance contrast and saturation of each
image. In addition, since a wide-angle
lens is used, edge distortion is corrected using a custom-built camera
model. Both of these overall algorithms
are executed automatically in a batch processing system.
The ASPECT camera is
fix-mounted to the primary optical base plate.
The camera axis is bore sighted to within 0.5 degrees to the axis
centers of the other optical systems.
While images are being collected, a concurrent system collects both GPS
data and inertial data to provide a high-resolution pitch, roll, and yaw
correction dataset. An automatic
software package merges these data set and geo-corrects each image using a
triangular correction mode. The
resulting images statistically show less than 11 meters of center frame
positional error and less than 1 degree of rotational error. As with the frame enhancement processing,
geo-registration is accomplished in a batch mode at a rate of approximately 800
images per hour. Following registration,
images can be directly used by the responder or further corrected with minor
positional and rotation corrections (Figure 26).
If requested by the data
user, aerial photography (and IR imagery) can be stitched into a wide area
mosaic. While this process does take
some time, a 4 square kilometer mosaic image (approximately 8 frames) can be
assembled in about 2 hours (Figure 27).
Oblique digital photography
is processed to capture the situational environment from the perspective of the
flight crew. All frames are collected
from the right side of the aircraft at approximately 45 degrees from the nose
of the aircraft. During automated
processing, GPS data is used to provide the position that the frame was
collected and the direction that the frame was collected is determined from the
track of the aircraft and the relative direction that the camera was operated
from within the aircraft. Figure 28
illustrates an example of an oblique image.
FIGURE 26 - Digital Aerial Visible
Imagery
FIGURE 27 -
Mosaic Imagery Product
FIGURE 28 -
Oblique Aerial Image
E.
Data Communication Technology
The ability to rapidly
transfer data from the ASPECT aircraft to the ultimate end user is mandatory if
the system is to support emergency response functions. ASPECT uses a state of the art
satellite-based communication system that provides broadband data through put
while the aircraft is in flight status (Figure 29). The system consists of an electronically
steered phase array satellite antenna coupled to a RF power amplifier/receiver
supporting a wired onboard computer TCP/IP modem/network. All components of the system have been
installed and certified as part of a formal FAA STC procedure. The system utilizes a geosynchronous
satellite connection and permits full rate communication throughout the
contiguous U.S. Table 11 contains the
technical specifications for the satellite communications system.
FIGURE 29 -
Satellite Communication System Phased Array Antenna
TABLE 11 -
Satellite Communication Technical Specifications
System: |
Chelton
Broadband Satellite System |
Antenna: |
HGA-7000
Electronically Steered Phased Array Antenna. |
Modem: |
Integrated
Airborne Modem/Router, 100 MB/s data rate |
Power
Amplifier: |
HPA-7400
Bi-directional Power Amplifier/Pre-Amplifier Short Coupled to the Phased
Array Antenna. |
Data Rate: |
Up to 332
kbs (Approximately 60 Kbs) Full Duplex
|
Constellation
Type: |
Fixed
Geo-Synchronous |
Coverage: |
Continuous
Coverage Over the Lower 48 States. |
Certification: |
FAA STC |
Power: |
28 vdc @ 10
amp full load |
Spin-up
Time: |
Less than 2
minutes from System Start |
Standards: |
TCP/IP |