Motion Compensation in ISAR Imaging Using Time-Frequency Techniques - Time-Frequency Transforms for Radar Imaging and Signal Analysis - page 133

Figure 6.4 (a) ISAR image of simulated Boeing-727 data after translation motion compen-

sation using one prominent point-scatterer; and (b) STFT spectrogram of the
signal at the chosen range cell. ( Source: [13] Ó 1998 IEEE.)

aligned range profiles over 128 pulses. Figure 6.7(b) shows the resulting
ISAR image from taking a series of 1D Fourier transforms across the pulse
number (or dwell time). Since significant phase errors still exist in the data,
the image is quite blurry in the Doppler dimension. Figure 6.7(c) shows
the spectrogram of the strongest range cell. It can be seen that the trajectories
of the point-scatterers are tilted. In addition, slight curvature can be noticed
in the trajectories. Therefore, we used polynomials of order three (i.e.,
including linear, quadratic, and cubic terms) in the basis search. We then
multiplied the radar data by the conjugate of the best-fit basis to remove