|
Collection of pushbroom sensor imagery from a mobile platform requires corrections using inertial measurement units
(IMU's) and DGPS in order to create useable imagery for environmental monitoring and surveillance of shorelines in
freshwater systems, coastal littoral zones and harbor areas. This paper describes a suite of imaging systems used during
collection of hyperspectral imagery in northern Florida panhandle and Gulf of Mexico airborne missions to detect
weathered oil in coastal littoral zones. Underlying concepts of pushbroom imagery, the needed corrections for directional
changes using DGPS and corrections for platform yaw, pitch, and roll using IMU data is described as well as the
development and application of optimal band and spectral regions associated with weathered oil. Pushbroom sensor and
frame camera data collected in response to the recent Gulf of Mexico oil spill disaster is presented as the scenario
documenting environmental monitoring and surveillance techniques using mobile sensing platforms. Data was acquired
during the months of February, March, April and May of 2011. The low altitude airborne systems include a temperature
stabilized hyperspectral imaging system capable of up to 1024 spectral channels and 1376 spatial across track pixels
flown from 3,000 to 4,500 feet altitudes. The hyperspectral imaging system is collocated with a full resolution high
definition video recorder for simultaneous HD video imagery, a 12.3 megapixel digital, a mapping camera using 9 inch
film types that yields scanned aerial imagery with approximately 22,200 by 22,200 pixel multispectral imagery (~255
megapixel RGB multispectral images in order to conduct for spectral-spatial sharpening of fused multispectral,
hyperspectral imagery. Two high spectral (252 channels) and radiometric sensitivity solid state spectrographs are used
for collecting upwelling radiance (sub-meter pixels) with downwelling irradiance fiber optic attachment. These sensors
are utilized for cross calibration and independent acquisition of ground or water reflectance signatures and for
calculation of the bi-directional reflectance distribution function (BRDF). Methods are demonstrated for selecting
optimal spectral regions and bands for discrimination, detection and characterization of weathered oil in the Northern
Gulf of Mexico waters and littoral zones in response to the Deepwater Horizon oil spill disaster. The techniques allow
for the use of sun and sky glint regions in imagery to identify water surface wave field characteristics as well as oil
slicks. The systems described provide unique data sets for remote sensing algorithm development and future testing of
radiative transfer models useful in studying weathered oil fate, distribution and extent.
|
