Signal Processing

Fully polarimetric E-SAR image of Oberpfaffenhofen, incl. DLR facilities in image center. Composite obtained from 2 dual-polarized
data takes in C-band (HH-green, HV-red, VV-blue)

Synthetic aperture radar systems usually require sophisticated signal processing treatment in order to allow the retrieval of the desired information from the data. Spatial resolution as well as radiometric and interferometric calibration accuracy have direct influence on the potential to measure or infer physical parameters.

The development of new algorithms to provide the required information and to improve the quality of the associated data products is an ongoing process. In addition the increasing amount of the acquired and processed E-SAR data poses strong requirements on efficient algorithm implementation. About 300-400 data takes per year where acquired during the last 4 years, corresponding in average to ca. 1GB of raw data per data take and also about 1 GB per processed data product.

E-SAR processor

The E-SAR processor includes:

  • the Extended Chirp Scaling (ECS) algorithm for image formation (high resolution, phase preserving, accurate motion error correction, no interpolations).
  • standard interferometric processing modules for single- and repeat-pass SAR interrferometry.
  • radiometric and polarimetric calibration.
  • geocoding modules for SAR image data and digital elevation models (DEMs).
  • segmentation & mosaic modules.

Linked to the requirements of high resolution and precise motion compensation together with the high number of flight campaigns abroad, the E-SAR processing software has been extensively adapted to run on individual LINUX-PC’s while assessing common data storage for the raw and processed data. During campaigns outside Europe (India, Indonesia, and Tunisia) this concept allowed on-site data processing and quality check.

E-SAR processing, data products and data archiving strategy
E-SAR data products

The E-SAR processor now provides the following operational products:

  • slant range polarimetric multilook and SLC data (Radar Geometry Images – RGI product).
  • Geocoded Terrain Corrected data (GTC product).
  • Digital Elevation Models (DEM product)
  • Repeat-Pass Interferometric data (RP-InSAR product, consisting of master & slave RGI products)

For archiving purposes, each data product is ingested into the Data Ingestion and Management System DIMS of the German Remote Sensing Data Center (DLR-DFD) allowing also world-wide access via EOWEB.

Recent Developments

  • DEM generation
  • Airborne Repeat-pass SAR Interferometry
  • Topographic motion compensation
  • Traffic Monitoring

Digital elevation model of Lago Maggiore area obtained from single-pass
interferometric E-SAR data in X-band (WGS-84, UTM zone 32)
DEM generation

Digital elevation models are one of the standard E-SAR products. However, large scale mapping is a challenge due to the small swath width of only 3,5 km. Therefore a mosaic procedure is used to generate large area DEMs. With the opportunity of data acquisition near Lago Maggiore (Italy) the mosaic procedure is extended to incorporate iterative topography adaptive compensation of residual motion errors. An elevation model covering an area of 15km by 15km was generated from 15 different tracks. Horizontal posting is 2m and height accuracy is better 2m rms (better 5m rms on steep slopes). Total height variation is about 1500m. Black areas correspond to lake areas (no signal information is available).

Airborne Repeat-pass SAR Interferometry

A dedicated processing chain for repeat-pass SAR interferometry was implemented which demands extremely high complexity in the processing chain. The accurate processing is a precondition to obtain high quality Pol-InSAR data products as input to the model based inversion approaches (more details Link to PolINSAR). The performance of standard airborne SAR processing is limited by the accuracy of the navigation data available to perform the motion compensation (state-of–the art is a combination of inertial and GPS sensors). Although the relative accuracy is very good, enabling well focused data, the absolute performance is limited by the absolute precision of the differential GPS signal which is in the order of 5-10cm (ca. one interferometric phase cycle assuming SAR data at L-band and the two way propagation delay). It is obvious that this accuracy is not sufficient for repeat-pass interferometry. A robust estimation approach was developed based on the so-called multi-squint technique. Residual errors (in horizontal and vertical directions) between the two tracks of the interferometric acquisitions are computed from the processed interferometric SAR data. An example is shown below:

Residual track errors measured and corrected by repeat-pass processing approach.
Almost perfect compensation is reached after 3-4 iterations, corresponding to a correction
of the residual motions errors with millimeter accuracy

The original residual track errors are in the order of 3-5cm and are in accordance with the accuracy of the navigation system based on differential GPS and INS. Without this compensation approach, phase errors and coherence degradations would occur which have strong impact on the interferomertric data quality.

Topographic motion compensation

During the development and refinement of the airborne repeat-pass processing strategy it was recognized that the usual approximations for motion compensation are not sufficient and that topography needs to be taken into account very precisely. This led to the development of a new algorithm, the so-called Precise Topography and Aperture dependent motion compensation (PTA) approach. It is based on short time FFT codes and thus makes effective use of the quasi linear azimuth time-frequency correspondence of the SAR signal. Its application is a pre-condition not only for processing airborne data in differential SAR interferometric mode, but also for repeat-pass SAR applications in hilly and mountainous areas in general.

Traffic Monitoring

The E-SAR system uses small antennas in order to avoid a gimble-based antenna steering configuration. This leads to an azimuth signal with a very high bandwidth associated with a correspondingly high PRF value.
During processing, usually only a small portion of the azimuth bandwidth (ca. 150Hz) is needed to obtain comparable resolutions in azimuth and range directions. The availability of a very high azimuth bandwidth (ca. 900 Hz in X-band) of the E-SAR system is beneficial for monitoring moving targets in the along-track interferometric mode, as the signal energy might be shifted outside the azimuth bandwidth used in the conventional processing. Therefore full azimuth bandwidth processing is required to allow MTI with sufficient accuracy. For this purpose, the E-SAR processor was extended to efficiently allow the allocation of enough memory and the generation of the corresponding phase filter functions. The result is a very high resolution image (up to 10cm in azimuth) and the associated along-track interferogram.
A part of such an image and the associated interferometric phase with a first indication of moving targets is shown below.

Full azimuth bandwidth E-SAR image in X-band (left) and corresponding along-track interferometric phase. Blue colour corresponds to zero phase and represents the stationary targets. The moving targets are shows in green, yellow and orange colours

Note that the resolution in range is limited by the bandwidth of the radar system of 100MHz and is by a factor of 15 worse than the resolution in azimuth. Further evaluation of these data sets is conducted as part of DLR’s TRAMRAD project and also within the TerraSAR-X traffic processor development .

Scientific topics under investigation

  • Airborne differential SAR interferometry
  • Airborne SAR tomography
  • Radar sounder data processing

Dr.-Ing. Rolf Scheiber
German Aerospace Center

Microwaves and Radar Institute
, SAR Technology
Tel: +49 8153 28-2319

Fax: +49 8153 28-1449

URL for this article