Desktop version

Home arrow Engineering arrow Study on Ground Moving Target Indication and Imaging Technique of Airborne SAR

Simulation Results and Analysis

In order to demonstrate the validity of the proposed algorithm, both simulation experiments and raw data are processed in this section.

In the simulation experiment, peak amplitude of the azimuth spectrum is set as 100, peak of Target A is 160, and peak of Target B is 40. To better simulate the robustness of the proposed algorithm, white Gaussian noise is added to the spectrum with amplitude of 5.

Figure 3.6a shows the simulated power spectrum of the azimuth data. Target A and B represent two kinds of prominent point targets. From Fig. 3.6b, the samples of Target A and B are far from fitting curve, and the peak of the fitting curve slightly departs from the actual Doppler center due to the impact of exceptional samples. Figure 3.6c shows that exceptional samples are abandoned, and new discrete azimuth spectrum is obtained. Figure 3.6d illustrates that after the second Gaussian curve fitting, the fitting curve is generally the same as the ideal azimuth power spectrum, and the center of the fitting curve is the same as the actual Doppler centroid. Therefore, the Doppler centroid can be estimated by operating traditional energy balancing algorithm. After processing the proposed algorithm, the homogeneity character of the scene is inherently improved, and the estimation precision is increased.

To testify the effectiveness of the proposed algorithm, imaging of real Ku-band airborne SAR data is processed. Figure 3.7a, b show focusing performance before and after proposed algorithm without autofocus technique. It is clear that Fig. 3.7b has a finer resolution, since Doppler centroid is estimated more precisely and RCM is completely corrected. Figure 3.7c offers the optical photo of the scene as a reference. Therefore, the validity of the proposed algorithm is confirmed.

Results of the simulation. a Simulated azimuth spectrum. b After the first Gaussian curve fitting. c Elimination of exceptional samples. d After the second Gaussian curve fitting

Fig. 3.6 Results of the simulation. a Simulated azimuth spectrum. b After the first Gaussian curve fitting. c Elimination of exceptional samples. d After the second Gaussian curve fitting

Results of raw data imaging. a Traditional algorithm. b Proposed algorithm. c Optical

Fig. 3.7 Results of raw data imaging. a Traditional algorithm. b Proposed algorithm. c Optical

image

 
Source
< Prev   CONTENTS   Source   Next >

Related topics