SAR Raw Data Imaging And Motion Compensation In The Absence Of Flight Data

National Aeronautics and Space Administration (NASA) Goddard Space Flight Center (GSFC) recently developed a state-of-the-art L-band airborne radar, named with "Digital Beamforming Synthetic Aperture Radar (DBSAR)", for the development and testing of digital beamforming radar techniques to earth and planetary measurements. To use multi-mode measurement techniques in a radar platform, DBSAR employs a number of advanced radar technologies, customized data acquisition and a real-time processor. The overall goal of the research is to process the DBSAR raw data into single look complex (SLC) images, and further consider motion compensation and interferometry applications, including generation of digital elevation models (DEMs), and dual-pass differential unwrapped phase maps.However, the high precision flight data associated with the DBSAR raw data are not available. The absence of the high precision flight data leads to some special challenges to DBSAR imaging in high image quality and phase preservation. This is because the motion parameters, including the forward velocity and near range distance, are indispensable for imaging algorithms. Further, for airborne synthetic aperture radar (SAR), the raw data are typically affected by the so-called motion errors due to the presence of atmospheric turbulence that produce sensor track deviations from an ideal straight track. For a better image quality, the forward velocity and near range distance must be firstly determined. Further, motion errors should be extracted and properly compensated. Within the scope of the thesis, a software system for airborne SAR raw data imaging and motion compensation is developed and investigated for the DBSAR raw data processing.For SAR imaging, also known as SAR focusing, the choice of a highly accurate and efficient approach is relevant to the quality and phase preservation of the focused image products. After comparisons of couples of imaging algorithm, such as the range-Doppler algorithm, chirp scaling algorithm, and the wavenumber domain algorithm, on the basis of several experiments, the chirp scaling algorithm is found to be efficient and high accuracy, and is thus chosen to focus the DBSAR raw data.To improve the image quality and phase preservation, sensor track must be properly corrected to the ideal straight track. The motion deviation information can be obtained from inertial navigation system (INS)/global positioning system (GPS) data, or raw data. Due to the absence of flight data for DBSAR, the motion information must be extracted from the raw data. To extract the motion parameters, a chirp-scaling-algorithm-based method is proposed and is employed to estimate the motion parameters. The proposed method for extraction of motion parameters is able to perfectly separate the motion parameters, i.e., the forward velocities in azimuth direction and the displacements in line-of-sight (LOS) direction. The main workflows of the proposed method consist of two steps. The first step is to estimate an instantaneous Doppler rate from radar raw data. The second step is to separate essential motion parameters, i.e., displacements and forward velocities in LOS direction, based on the estimated Doppler rate in the first step.For implementation of the motion compensation, a chirp-scaling-algorithm-based method is proposed. The motion compensation is separately implemented in range direction and azimuth direction. The motion compensation is firstly implemented in range direction by compensating the phase caused by the displacements in LOS direction and shifting the corresponding envelopes in range-Doppler domain. Then, the motion compensation is implemented in azimuth direction and in the range-Doppler domain by subtracting the Doppler rate error phase from the actual phase history and resampling the focused image to compensate the azimuth position deviation after azimuth compression. The proposed chirp scaling algorithm integrated with the motion compensation strategy yields a superior image quality as compared to the traditional chirp scaling approach that use the uniform velocity and uniform near range distance, regardless of the variation of forward velocities and displacements.Another emphasis of this thesis is validation of the accuracy level of phase preservation of the SLC images focused using the parameters estimated from raw data via implementation of interferometry. Due to the absence of high precision flight data, it is impossible to carry out extremely accurate motion compensation. The phase of the focused SLC may not be preserved accurately. Further, we only have a scene of DBSAR raw data; thereby, the author is unable to perform a repeat pass interferometry on the provided DBSAR data. To prepare the next mission, where the high precision flight data would be ready for accurate motion compensation and two- or multi-pass interferometry would be carried out, the author performs a comprehensive assessment scheme for generation of digital elevation model (DEMs) and dual-pass differential unwrapped phase maps using parameters estimated from a pair of Advanced Land Observing Satellite/Phased Array type L-band Synthetic Aperture Radar (ALOS/PALSAR) raw data. The primary experiments showed that for spaceborne SAR it is possible to the parameters estimated from raw data to create SLC, DEM and Dual-pass differential phase maps when accuracy requirements are not very high. In other words, the accuracy level of phase preservation of the SLC images focused using the parameters estimated from raw data is good enough for implementation of interferometry. This is another clear indicator of the effectiveness of the proposed motion compensation method in the absence of flight data. This attempt provides a positive reference for implementation of interferometry in the next DBSAR mission. These focused SLC images will help other people in NASA/GSFC measure the three dimensional structures (in terms of tree height and density) of North American forests that is however beyond the scope of this study.