Study Of Several Problems In SAR Imaging

Synthetic aperture radar has evolved to satisfy a variety of applications for both civilian and military users with its fine two梔imensional resolution capability and possibility of day/night operation in all weather. This dissertation investigates the problem for synthetic aperture radar imaging algorithms and their application, focusing on Doppler parameter estimation algorithm, radar imaging algorithms with high squint and the applications and performance analysis of radar imaging algorithms. The main contents oI the dissertation are described as following. Firstly, a brief introduction of the purpose. background and significance of the ~vork~ in the dissertation is concerned. The azimuth power spectra may be distorted due to the presence of the bright targets that has been illuminated by a partial aperture, thus degrading the estimation precision w the Doppler centroid frequency. In chapter 1, we propose a time-frequency filtering method based on the analysis of the model of time-frequency distribution of the Doppler phase history. This method can improve the azimuth power spectra and the precision of Doppler centroid frequency estimation by filtering the time-frequency distribution of the tars.~cI illttrninated by partial aperture. With introduction the entropy to measure the extent of the image focusing and use of it to estimate a Doppler frequency rate, a new technique I鑢 estimating the Doppler frequency rate is proposed which requires a small amount ot computation and gives a high accuracy. ERS- 1 raw data are used for imaging test by the use of Doppler parameters obtained with Doppler parameter estimation methods proposed in this chapter, and well-focused images are generated. Based on the analysis of the radar echo with high squint, a separable ima~in~ approach for processing high squint mode SAR data is proposed in chapter 3. Tli~s approach involves a time domain range walk compensation, high order Doppler signal approximation for high squint mode, partitioning processing along range for point scatterer go through focus depth causing by time domain range walk compensation. Finally the approach is demonstrated efficiently by computer simulation results. Chapter 4 analyzes the specific property of E-SAR data. A new imaging algorithm combined with the Chirp-Scaling technique is proposed, which instead of performing geometric correction individually, corrects range cell migration and geometric distortion .Iv Abstract simultaneously via Chirp Scaling operation. The focusing in azimuth is done by use of the de-Chirping technique. This method is suitable for imaging using partial aperture data. The comparison of the proposed algorithm with the quick look imaging algorithm which is based on the de-Chirping technique shows that the performance of the former is much better. Reference frequency is an important parameter in nonlinear chirp scaling algorithm. It will degrade the performance of the algorithm if reference frequency is selected mistakenly. Based on analyzing the nonlinear chirp scaling algorithm and the factors that affect reference frequency selection thoroughly, the criterion to select reference frequency is determined in Chapter 5. In addition, radar image is distorted along range at a large squint angle observation, Multiple-look SAR imaging method for a large squint angle. is presented which compensates the range distortion. The phase approximation can cause phase...