G51C-01
The UAVSAR Instrument: Description and First Results
The UAVSAR instrument, employing an L-band actively electronically scanned antenna, had its genesis in the ESTO Instrument Incubator Program and, after 3 years of development, has begun collecting engineering and science data. System design was motivated by solid Earth applications where repeat pass radar interferometry can be used to measure subtle deformation of the surface. However, flexibility and extensibility to support other applications were also major design drivers. The radar is designed to be housed in an external unpressurized pod and has the potential to be readily ported to many platforms. Initial testing is being carried out with the NASA Gulfstream III aircraft, which has been modified to accommodate the radar pod and has been equipped with precision autopilot capability developed by NASA Dryden Flight Research Center. With this, the aircraft can fly within a 10 m diameter tube on any specified trajectory necessary for repeat-pass radar interferometric applications. To maintain the required pointing for repeat-pass interferometric applications, we have employed an actively scanned antenna steered using INU measurement data. We will present a brief overview of the radar instrument and some of the first imagery obtained from the system.
G51C-02 INVITED
Geodetic imaging with time series persistent scatterer InSAR
Measuring the temporal evolution of deformation is crucial for understanding geophysical processes. InSAR methods produce detailed, spatially dense geodetic images in areas with high correlation between scattered echoes on different radar passes. Decorrelation occurs when the surface changes at wavelength scale over time. Persistent scattering (PS) methods are one way to avoid much observed decorrelation, and have proven to be useful in extracting deformation signatures from some highly decorrelating natural terrains. For slowly varying deformation fields, such as fault creep, PS approaches can image the evolution of surface movements. For episodic events such as volcanic events or even variable subsidence applications, PS networks often underestimate the deformation. The limitations of the technique appear to be mainly related to the sparseness of the data in both space and time, and are directly related to the robustness of the phase unwrapping algorithm needed to connect deformation points spatially and temporally. While early PS implementations were most successful in examining urban areas with manmade structures, and hence many bright backscattering points, the development of phase-based persistent scatterer identification approaches has permitted the method to be applicable for geophysical research. We now use a maximum likelihood phase-based method for PS identification that can find even low-amplitude PS points, greatly improving the spatial density of reliable points over that of earlier phase coherence approaches. We have applied the new detection methods to several different terrain types for various applications and present here sample PS network interferograms. We have verified several of these using either GPS or leveling data and show that the PS solutions are sometimes overly smoothed temporally, especially where the data are still poorly sampled in time. The smoothing is an artifact of our inability to properly unwrap data in three dimensions. We present here examples of methods we are developing to unwrap these data, so that PS analysis can successfully retrieve time series deformation records. Of particular promise are fusion methods using principles of SBAS reduction to constrain the PS phase unwrapping
G51C-03
Modelling of deformations occurring in the city of Auckland, New Zealand and observed by Differential Synthetic Aperture Radar
Auckland is the largest city in New Zealand with a current population of over one million. It is situated on a basaltic volcanic field which consist of over 50 individual largely monogenetic volcanoes with a total area of 360 sq. km. The most recent and largest eruption occurred 600 years ago, and was witnessed by local inhabitants. It is anticipated that the chance of reawakening of a past volcano is very low; however, a new volcano could be created at any time in a new location within the field. In this work we present results of modelling of the deformations that occurred in the city of Auckland from 18 July 2003 to 9 November 2007. These deformations were observed by the Differential Synthetic Aperture Radar on ENVISAT satellite (Track 151, Frame 6442, IS2, VV). Stacking, Small Baseline Subset (SBAS) and Permanent Scatterers (PS) processing algorithms where used to determine spatial and temporal patterns of surface deformation as well as average rates. A number of localized deformation regions were consistently observed by all three techniques. Three regions of subsidence are believed to be caused by groundwater extraction. And three source of uplift are modeled here as volcanic sources, however, the volcanic nature of these uplifts has not been confirmed.
G51C-04
Imaging the Ionosphere Using Polarimetric SAR and GPS
In this paper, we will present ionospheric electron content measurements derived from polarimetric data collected using the Phased Array type L-band Synthetic Aperture Radar (PALSAR) onboard the Japanese Advanced Land Observing Satellite (ALOS). The radar measurements of latitudinal features in different regions under various conditions are compared with 2D and 3D assimilative ionsopheric models that ingest GPS data. The comparisons show consistent results between the SAR- and GPS-based techniques. This indicates that GPS-based techniques could be applied to corrections of SAR data, at least to the first order approximation, that are subject to the Faraday rotation and other ionospheric effects. In addition, while GPS measurements collected from global and regional GPS networks are useful to provide large-scale snapshots of ionospheric TEC maps and 3D densities, polarimetric SAR could become a powerful tool to image the ionosphere in much higher resolutions. The potential of SAR-based ionospheric imaging will be discussed that can significantly enhance our capability of investigating outstanding ionospheric research topics.
G51C-05 INVITED
Estimating and Improving the Accuracy of Airborne Kinematic GPS Positioning
Accurate airborne LIDAR mapping is critically dependent on accurate positioning of the aircraft's trajectory. This positioning is done using either GPS by itself or in conjunction with INS. This paper will examine the accuracies with which GPS can determine airplane positions. This is customarily done by using several ground-based reference stations; to each of which a separate differential, kinematic GPS solution is computed. These solutions use the L1 and L2 carrier phases to form the double differenced, ionosphere-free observable with the phase bias's fixed to their integer values. Precise GPS orbits computed by the IGS are also used. Differences between the trajectories computed using several base stations will reveal the effects contributed from those base stations, e.g. multipath, position errors, etc. Common mode errors arising from the aircraft however, will not be seen. The KinTools program has been used to compare trajectories for a number of NCALM mapping projects. These projects, which used base stations distributed over 1500m in elevation and with distances to the airplane ranging from 0 to 250km, have shown the importance of estimating tropospheric delay as a function of airplane altitude. This estimation requires that the GPS antenna on the aircraft be calibrated for phase center offsets and phase center variations. The kinematic data itself is then used to model the variation of zenith neutral delay. The resulting trajectory RMS vertical agreement is typically better than 2cm and is independent of the range to the base stations, the base station elevation and the airplane altitude.
G51C-06 INVITED
Revealing Patterns in Shoreline Change with High-Resolution Airborne Lidar
Data mining and pattern classification techniques offer great potential to move the analysis of coastal lidar data beyond visual interpretation and simple (first order) measurements made from digital elevation models (DEMs). This is particularly true for coastal monitoring with lidar data because of the importance of small- scale features in the DEMs. When acquired with sufficient temporal coverage, the high spatial-resolution information in the lidar data can reveal patterns in beach change. Here, we develop methods to characterize shoreline change response to variation in beach morphology. Results are based on airborne lidar data acquired along a beach in Florida multiple times between August 2003 and March 2008. Several morphologic features are extracted and then segmented into binary shoreline change classes. Binary regression models are developed to "learn" the influence of the features on the probability of eroding or accreting. Those features more indicative of the observed shoreline change patterns are systematically detected providing insight into the change dynamics. The developed models can be used for classification of high-impact zones and several examples are presented.
G51C-07
Generating Precise and Accurate Waveform-Derived Products From Medium-Footprint Geodetic Imaging Lidar
NASA's airborne Laser Vegetation Imaging Sensor (LVIS) is a medium-high altitude (10 km above the ground), medium-footprint (10-25 m wide) wide-swath (2 km) geodetic, imaging laser altimeter system that digitally records the shapes of both the outgoing and returning laser pulses (waveforms) for every shot. Since 1997, the system has been used to acquire data in various locations including California, Arkansas, Costa Rica, New England, Maryland, Virginia and Greenland. It has also been used to prototype future spaceborne measurements (e.g., VCL, DESDynI), develop and refine data and return waveform processing algorithms, and showcase science applications of full-waveform altimetry. Data geolocation is achieved using NASA's Variable Estimation, Geolocation and Analysis Software (VEGAS). Using this software, we solve for and apply various system biases and parameters (e.g., angular offsets between the various reference frames on the aircraft and the timing offsets between the various data inputs) to generate the position of the laser footprint on the ground relative to a global reference frame such as WGS84. Waveforms are interpreted to identify the locations of different surfaces within the footprint including the mean ground, canopy top, and several metrics related to surface (including canopy) structure. For each data set, the precision and accuracy of the data products are assessed by intercomparison, and comparison to available in-situ data such as collected using GPS or higher-resolution laser altimetry. We will review the latest data processing procedures employed in the VEGAS system and the LVIS waveform interpretation approaches. The precision and accuracy of the data products from various locations will also be presented. Implications for future spaceborne missions recommended by the National Research Council (NRC) such as DESDynI, ICESat II and LIST will be discussed.
G51C-08
Topographic and Bathymetric Surface Mapping Using Low-SNR Airborne Lidar
Recent technological advances in the performance of small micro-lasers (sub-nanosecond pulses with energies of a few micro Joules) and multi-channel, multi-event photo-detectors have enabled the development of experimental airborne lidar systems based on a low SNR paradigm. This emerging class of lidar sensors allows for very dense point spacing on the ground and vertically through vegetation. As a result, these systems offer the potential to significantly increase the fidelity of terrain reconstruction over what is currently possible with conventional lidars that employ pulses that are several nanoseconds long with energies of roughly one hundred micro Joules. The low-SNR systems are, however, highly sensitive to background sources of light because they rely on the detection of events on the order of a single photon. A numerical sensor simulator has been developed to model the expected output from low-SNR airborne lidar systems and predict their performance. Although availability of data from actual low-SNR lidar sensors is still very limited, prototypes developed at NASA and the University of Florida offer some experimental results that are used to validate the simulation. The sheer volume of data returns, in combination with the increased number of noise returns, introduces complications to data processing and analysis. A filtering algorithm is presented that demonstrates the ability to automatically remove a majority of background noise. The topographic surface is then reconstructed and compared to truth to demonstrate the capability of low-SNR lidar systems to observe small-scale elevation change. Resolving depth in shallow water using current bathymetric lidar systems remains problematic because of difficulties in distinguishing surface and bottom returns due to the large transmitted pulse widths necessitated by high pulse energies. The low transmit pulse energies of low-SNR lidar systems offer a promising advantage in shallow waters because of the sub-nanosecond pulse widths. Simulated data is presented that shows low-SNR lidar's potential advantage in detecting shallow stream bottoms. Modifications to the filtering algorithm are described that allow for extraction of stream bottoms and surface returns, allowing for geometric correction due to index of refraction effects.
G51C-09
Assessment of the Geodetic and Color Accuracy of Multi-Pass Airborne/Mobile Lidar Data
The ability to merge lidar and color image data acquired by multiple passes of an aircraft or van is largely dependent on the accuracy of the navigation system that estimates the dynamic position and orientation of the sensor. We report an assessment of the performance of a Riegl Q560 lidar transceiver combined with a Litton LN-200 inertial measurement unit (IMU) based NovAtel SPAN GPS/IMU system and a Panasonic HD Video Camera system. Several techniques are reported that were used to maximize the performance of the GPS/IMU system in generating precisely merged point clouds. The airborne data used included eight flight lines all overflying the same building on the campus at Utah State University. These lines were flown at the FAA minimum altitude of 1000 feet for fixed-wing aircraft. The mobile data was then acquired with the same system mounted to look sideways out of a van several months later. The van was driven around the same building at variable speed in order to avoid pedestrians. An absolute accuracy of about 6 cm and a relative accuracy of less than 2.5 cm one-sigma are documented for the merged data. Several techniques are also reported for merging of the color video data stream with the lidar point cloud. A technique for back-projecting and burning lidar points within the video stream enables the verification of co-boresighting accuracy. The resulting pixel-level alignment is accurate with within the size of a lidar footprint. The techniques described in this paper enable the display of high-resolution colored points of high detail and color clarity.