Interferometric Synthetic Aperture Radar - Producing Interferograms

Producing Interferograms

The processing chain used to produce interferograms varies according to the software used and the precise application, but will usually include some combination of the following steps.

Two SAR images are required to produce an interferogram; these may be obtained pre-processed, or produced from raw data by the user prior to InSAR processing. The two images must first be co-registered, using a correlation procedure to find the offset and difference in geometry between the two amplitude images. One SAR image is then re-sampled to match the geometry of the other, meaning each pixel represents the same ground area in both images. The interferogram is then formed by cross-multiplication of each pixel in the two images, and the interferometric phase due to the curvature of the Earth is removed, a process referred to as flattening. For deformation applications a DEM can be used in conjunction with the baseline data to simulate the contribution of the topography to the interferometric phase, this can then be removed from the interferogram.

Once the basic interferogram has been produced, it is commonly filtered using an adaptive power-spectrum filter to amplify the phase signal. For most quantitative applications the consecutive fringes present in the interferogram will then have to be unwrapped, which involves interpolating over the 0 to 2π phase jumps to produce a continuous deformation field. At some point, before or after unwrapping, incoherent areas of the image may be masked out. The final processing stage involves geocoding the image, which resamples the interferogram from the acquisition geometry (related to direction of satellite path) into the desired geographic projection.