Abstract

Remote sensing radar satellites cover wide areas and provide spatially dense measurements, with millions of scatterers. Knowledge of the precise position of each radar scatterer is essential to identify the corresponding object and interpret the estimated deformation. The absolute position accuracy of synthetic aperture radar (SAR) scatterers in a 2D radar coordinate system, after compensating for atmosphere and tidal effects, is in the order of centimeters for TerraSAR-X (TSX) spotlight images. However, the absolute positioning in 3D and its quality description are not well known. Here, we exploit time-series interferometric SAR to enhance the positioning capability in three dimensions. The 3D positioning precision is parameterized by a variance–covariance matrix and visualized as an error ellipsoid centered at the estimated position. The intersection of the error ellipsoid with objects in the field is exploited to link radar scatterers to real-world objects. We demonstrate the estimation of scatterer position and its quality using 20 months of TSX stripmap acquisitions over Delft, the Netherlands. Using trihedral corner reflectors (CR) for validation, the accuracy of absolute positioning in 2D is about 7 cm. In 3D, an absolute accuracy of up to ?66 cm is realized, with a cigar-shaped error ellipsoid having centimeter precision in azimuth and range dimensions, and elongated in cross-range dimension with a precision in the order of meters (the ratio of the ellipsoid axis lengths is 1/3/213, respectively). The CR absolute 3D position, along with the associated error ellipsoid, is found to be accurate and agree with the ground truth position at a 99% confidence level. For other non-CR coherent scatterers, the error ellipsoid concept is validated using 3D building models. In both cases, the error ellipsoid not only serves as a quality descriptor, but can also help to associate radar scatterers to real-world objects.

Highlights

  • Interferometric synthetic aperture radar (InSAR) has evolved towards an effective tool for measuring the Earth’s topography and surface deformation

  • corner reflectors (CR) phase center positions measured with differential global positioning system (DGPS) and tachymetry were radar-coded and compared with image pixels which were fast Fourier transform (FFT) oversampled by a factor of 128 × 128

  • The secondary positioning components such as solid earth tides (SET), azimuth time shift, path delay and plate tectonics were computed as shown in Fig. 7, and corrected to improve the absolute positioning of CR6 and CR7

Read more

Summary

Introduction

Interferometric synthetic aperture radar (InSAR) has evolved towards an effective tool for measuring the Earth’s topography and surface deformation. Persistent scatterer interferometry (PSI) is one of the techniques to process a set of images in order to identify phase-coherent scatterers known as persistent scatterers (PS) (Ferretti et al 2001; Kampes 2005). The relative displacement is estimated with millimeter-level precision, but the positioning precision is usually in the order of decimeters or even meters. This hampers the interpretation of the deformation signal, as it is unclear which object is associated with the measurements. The absolute position accuracy of ENVISAT (Small et al 2004a, b, 2007) and Sentinel-1A (Schubert et al 2015) images were computed to be in the order of several decimeters at best both in azimuth and range directions.

Scatterer positioning
The dimensionless 2D radar datum
Quality description
Scatterer position validation
Configuration
Computation of the secondary positioning components
Results
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call