ALBERT

Description: The advantages and disadvantages of three different approaches to solving the problem of the radiometric correction of synthetic aperture radar (SAR) images of varying terrain heights are presented. The first approach involves registration of a digital elevation model (DEM) of the terrain to the image, determination of the local elevation and incidence angles, and appropriate radiometric correction. The second approach uses a DEM generated from interferometric SAR data to derive the elevation and incidence angle maps. In the third approach, a monopulse technique is employed to determine the elevation angle only. The relative errors in radiometric correction between these approaches are assessed. Calibration errors are estimated using corner reflectors deployed within some of the scenes imaged by the Jet Propulsion Laboratory airborne SAR (JPL AIRSAR).

Description: The POLCAL crosstalk removal algorithm is based on the statistical properties of the image background and does not need any corner reflector or active radar calibrator deployed in the scene. The advantage of being able to remove the crosstalk contamination without using external calibration targets gives rise to the consideration of algorithms based on clutter statistics in polarimetric synthetic aperture radar (SAR) images in more detail. The basic assumption of the POLCAL procedure is the decorrelation of copolarized and cross-polarized backscatter, which is valid for natural targets with azimuthal symmetry. This assumption and others regarding the properties of the imaged surface are examined. An improved version called POLCALII is presented, and its performance is compared with POLCAL.

Description: Terrain height variations in mountainous areas cause problems in radiometric corrections of synthetic aperture radar (SAR) images. To determine the elevation angle and the height at the different parts of an image, an application of the monopulse principle is proposed. From the ratios of images radiometrically modulated by the difference and sum antenna pattern in range it is possible to calculate the appropriate elevation angle at any point in the image. Design considerations for a corresponding airborne SAR-system are presented, and some estimates of error influences (e.g., ambiguities), expected performance and precision in topographic mapping are given.

Description: Terrain height variations in mountainous areas cause two problems in the radiometric correction of SAR images: the first being that the wrong elevation angle may be used in correcting for the radiometric variation of the antenna pattern; the second that the local incidence angle used in correcting the projection of the pixel area from slant range to ground range coordinates may vary from that given by the flat earth assumption. We propose a novel design of a SAR system which exploits the monopulse principle to determine the elevation angle and thus the height at the different parts of the image. The key element of such a phase monopulse system is an antenna, which can be divided into a lower and upper half in elevation using a monopulse comparator. In addition to the usual sum pattern, the elevation difference pattern can be generated by a -pi phase shift on one half of the antenna. From the ratios of images radiometrically modulated by the difference and sum antenna pattern in cross-track direction, we can derive the appropriate elevation angle at any point in the image. Together with the slant range we can calculate the height of the platform above this point using information on the antenna pointing and the platform attitude. This operation, repeated at many locations throughout the image, allows us to build up a topographic map of the height of the aircraft above each location. Inversion of this map, using the precisely determined aircraft altitude and the accurate flight path, leads to the actual topography of the imaged surface. The precise elevation of one point in the image could also be used to convert the height map to a topographic map. In this paper, we present design considerations for a corresponding airborne SAR system in X-Band and give estimates of the error due to system noise and azimuth ambiguities as well as the expected performance and precision in topographic mapping.

Description: Fine-structure splittings of atomic oxygen (O-16) in the ground state have been accurately measured using a tunable far-infrared spectrometer. The 3P0-3pl splitting is 2,060,069.09 (10) MHz, and the 3Pl-3P2 splitting is 4,744,777.49 (16) MHz. These frequencies are important for measuring atomic oxygen concentration in earth's atmosphere and the interstellar medium.

Description: Using tunable far-infrared spectrometers with high-frequency stability and accuracy, the self-pressure broadening and shift of CH3CN are measured. Evidence of absorber-perturber resonance effects on the collisional line shape are obtained. This tests the theoretical model and its possible improvements and also allows predictions of broadening and shift for a large class of molecules. Moreover, the resonance effect produces a theoretical temperature dependence of self-broadening that is different from what is commonly assumed.

Description: The frequency of the astronomically important J = 1-0 rotational transition of HD at 2.7 THz (90/cm) has been measured with tunable FIR radiation with an accuracy of 150 kHz. This frequency is now known to sufficient accuracy for use in future astrophysical heterodyne observations of HD in planetary atmospheres (reported by Bezard et al., 1986) and in the interstellar medium (reported by Bussoletti et al., 1975).

Description: Rotational and fine-structure transitions between the low rotational levels of the OH radical in its X 2Pi state have been observed in absorption in the laboratory. It has thus been possible to measure the frequencies of these transitions directly. The observations were made with tunable far-infrared radiation generated by mixing two chosen CO2 laser frequencies in a metal-insulator-metal diode; the far-infrared difference frequency was radiated from the diode's whisker antenna. The measurements have an accuracy of a few hundred kHz. They both confirm and improve on the best previous estimates, which were obtained by extrapolation of laser magnetic resonance data.