Synthetic Aperture Radar (SAR) is an active microwave sensor technique that has the ability to produce high-resolution SAR images of terrain. A moving platform transmits series of coherent radar chirp pulses directed sideways of the flight trajectory. Coherent addition of the echoes received from the illuminated area yields so-called 2D SAR raw data, from which, after SAR processing, 2D SAR image data are obtained.
SAR, unlike optical systems, is an active radar system and hence works in all weather conditions and at night. Therefore it has been widely applied in remote sensing and mapping.
For transmission of airborne SAR data from a platform to a receiving ground station the amount of SAR data has to be reduced to not exceed the available channel capacity. The interpretability of the SAR images at the ground station, at least within regions of interest, should not be impaired by the data rate reduction.
First investigations have shown that the processing of SAR raw data to SAR image data must take place on board the platform, in order to stay within the channel capacity limit. On-board processing therefore comprises SAR processing resulting in focused SAR image data as well as data compression. The compressed bitstream can be sent to a ground station and can be decoded into focused SAR reference images, such images which can directly be evaluated by human experts. Correspondingly, the investigated system comprises two tasks:
1. Generating well-focused SAR image data automatically
2. Efficient coding of the SAR image data
3. Usability-Based SAR System Assessment:
1. SAR imaging assumes that the platform, which can be an airplane or a satellite, moves on a straight line at constant velocity. Real flight paths differ from that assumption which, if uncompensated, create image artifacts. Compensation based on motion information is conceptually possible. But motion parameters as obtainable from on-board GPS (Global Positioning System) and INS (Inertial Navigation System) often lack accuracy, so that so called autofocus algorithms have to be used in order to obtain well-focused images. Our autofocus approach is based on the Phase-Gradient-Algorithm, which estimates and corrects the phase error produced by the residual motion errors.
2. For an efficient coding of the focused SAR image data, a pre-processing has been developed, which dynamically converts the focused SAR image data into reference image data. Current research focuses on the data rate reduction required to transmit the reference image data within the limited channel capacity and without any loss of important image content. Because of the similarity of the reference SAR images to electro-optical images, the investigated SAR-codec is based on a standardised image codec extended by SAR specific reduction techniques.
3. Human evaluation of SAR images as an indirect means for usability-based assessment of the SAR systems from which the images arose, is time-consuming and costly. An assessment system is developed which aims at finding digital signal processing algorithms to simulate, complement and partly replace the human evaluation of SAR images.