Multiple Moving Target Indication and Spectrum Extraction
According to the analysis in Sect. 2.3.2, moving targets are classified into three categories by the relationship between its spectrum and the spectrum of the clutter. In real SAR signal processing, the parameters of the moving target is unknown. Therefore, it is rigorous to assume that all three kinds of targets exist in the scene. In order to indicate targets of different kinds, new algorithms must be proposed.
The Two-Step GMTI Algorithm
The principle of the GMTI algorithm for SAR is to suppress the energy of the clutter, and then indicate the moving target according to the result of the amplitude judgment, such as the CFAR detection. Therefore, the main aim of the GMTI algorithm is to suppress the clutter, and raise the SCR of the moving target.
According to Sect. 2.3.2, the moving targets with different motion parameters belong to different types. The targets of Type I are located at the high-band of the PRF, and the clutter are mostly located at the low-band. As a result, the targets of Type I have a high SCR in the range-Doppler domain. On the contrary, targets of Type II and III are submerged by the clutter. The SCR of the targets of Type II and III are low in the range-Doppler domain. After the azimuth compression, the energy of the moving targets of Type II and III are concentrated in the image domain, and the energy of the clutter is distributed in the whole image. If the energy of the moving target is strong, it can be indicated in the image domain. Therefore, a two-step GMTI algorithm is proposed to detect the moving targets of all the three types, and its specific processing steps are shown as follows.
Step 1: Perform the range compression on the echo, and transform it into the range-Doppler domain. In the range-Doppler domain, the spectra of the moving targets of Type I are located at the high-band of the PRF. Extract the signal of the high-band of the PRF using a high-pass filter. If there are multiple trajectories at the high-band of the PRF, each trajectory can be regarded as a moving target.
Perform the azimuth compression on the extracted signal, and detect the moving targets by the CFAR technique. After the detection, the energy of the moving targets should be eliminated from the clutter.
Step 2: After Step 1, the moving targets of Type I have been indicated and extracted. Divide the whole synthetic aperture into two sub-apertures in the range-Doppler domain, and perform the azimuth compression on both sub-aperture data. Two sub-aperture images can be obtained, in which the moving targets have different Doppler histories while the clutter has the same. Therefore, the clutter can be suppressed by the complex subtraction of the two sub-images, and the moving targets can be indicated in the image domain. The targets of Type II and III can be indicated in this step, only if their energy is strong enough to be distinguished after the clutter suppression.
The whole GMTI algorithm is shown above. The reason why the proposed GMTI algorithm uses a two-step indication is that: in real GMTI applications, multiple moving targets with different motion characters must be treated separately. As to the weak targets with fast cross-track velocities (Type I), it is unable to indicate them in the image domain. One the contrary, the image domain algorithm is highly suitable for the indication of the strong targets with slow cross-track velocities and the along-track moving targets.
It is also worth noticing that the proposed algorithm is still unable to indicate the slow and weak moving targets. This is the inherent GMTI limitation of the single-antenna SAR system. In fact, it is still a challenging issue even for the multi-channel SAR system in real applications.