In optical sensors, the features can be seen in the raw data with some radiometric and geometric distortions but SAR raw data are unfocused images where no feature can be seen in the image. Unlike optical sensors, signal processing algorithms are required to focus the SAR raw data. The processing of SAR raw data is always a challenge because it is highly dependent on SAR acquisition mode, processing parameters, and satellite state vectors. SAR acquires the data by transmitting the Linear Frequency Modulated (LFM) signal or chirp towards the Earth’s surface. The chirp has varying frequency and match-filtering techniques are used to compress the chirp at the desired resolution.

Figure 1: EOS-4 SAR Raw Data processing (Area: Dehradun)

The transmitted chirp helps to obtain the desired resolution in the range direction of the SAR image. The chirp interacts with different features on the ground and some of the signals are backscattered towards the SAR sensor. The SAR collects the backscattered chirp signal through the receiver. However, in the azimuth direction, the motion of the SAR will form the LFM signal due to the Doppler effect. Thus, the desired resolution in the azimuth direction will be achieved through the LFM signal generated through the Doppler effect. Figure 1(a) shows the chirp patterns in the received EOS-4 SAR raw data which is due to the strong double bounce scattering. Figure 1(b) shows the processed EOS-4 SAR image of Dehradun, Uttarakhand, India. The processed SAR image shows the urban features as bright pixels due to double bounce scattering.

Figure 2 : EOS-4 SAR Raw Data processing (Area: Mumbai)

Figure 2 (a) & (b) shows the full scene of raw & processed EOS-4 SAR image of Mumbai, India. Figure 2 (c) shows the processed EOS-4 image at an actual resolution of 3 m.

In SAR raw data processing, match filtering techniques are applied in both range and azimuth direction in the frequency domain. In the match filtering technique, the transmitted chirp is multiplied by complex conjugate of the received SAR signal in the frequency domain. These match-filtering techniques are called range compression and azimuth compression in the SAR raw data processing. In range compression, match filtering is applied in the range direction, and in azimuth compression, match filtering is applied in the azimuth direction. The other processing steps in the SAR raw data processing will include Range cell migration correction (RCMC), Doppler centroid estimation, etc. The complexity of SAR raw data processing also depends on the acquisition mode of the SAR. The most common mode of acquisition of SAR data is Stripmap mode where Range- doppler algorithm or chirp scaling algorithm can be used to focus the SAR image.

References:

1) Charles Elachi, Spacenorne RADAR remote Sensing Application And Techniques, IEEE press

2) G.Carrara, R.S.Goodman, and R.M.Majewski, Spotlight Synthetic Aperture Radar, Norwood, MA: Artech House

3) G. Franceschetti, R. Lanari, Synthetic Aperture Radar Processing, CRC Press, Boca Raton, F.L.,1999.

4) J.C. Curlander and R.N. McDonough, Synthetic Aperture Radar Systems and Signal Processing, New York, Wiley, 1991

5) Fawwaz T. Ulaby and M. C. Dobson, Radar Scattering Statistics for Terrain, Artech House.