Autofocus algorithm for curvilinear SAR imaging
2012
We describe an approach to autofocusing for large apertures on curved SAR trajectories. It is a phase-gradient
type method in which phase corrections compensating trajectory perturbations are estimated not directly from
the image itself, but rather on the basis of partial" SAR data { functions of the slow and fast times { recon-
structed (by an appropriate forward-projection procedure) from windowed scene patches, of sizes comparable to
distances between distinct targets or localized features of the scene. The resulting partial data" can be shown
to contain the same information on the phase perturbations as that in the original data, provided the frequencies
of the perturbations do not exceed a quantity proportional to the patch size.
The algorithm uses as input a sequence of conventional scene images based on moderate-size subapertures
constituting the full aperture for which the phase corrections are to be determined. The subaperture images
are formed with pixel sizes comparable to the range resolution which, for the optimal subaperture size, should
be also approximately equal the cross-range resolution. The method does not restrict the size or shape of the
synthetic aperture and can be incorporated in the data collection process in persistent sensing scenarios.
The algorithm has been tested on the publicly available set of GOTCHA data, intentionally corrupted by
random-walk-type trajectory
uctuations (a possible model of errors caused by imprecise inertial navigation
system readings) of maximum frequencies compatible with the selected patch size. It was able to eciently
remove image corruption for apertures of sizes up to 360 degrees.
Keywords:
- Correction
- Source
- Cite
- Save
- Machine Reading By IdeaReader
0
References
0
Citations
NaN
KQI