SLC to GRD with phases


i would like to generate GRD products but keeping the phase. I am not sure this is possible. I tried to use the S1 SLC to GRD tool, select option “save as complex output” but i get the following error : “Error nodeid Calibration: Please select I and Q bands in pairs only”


When converting SLC product into GRD, phase is not retained. If you want to perform interferometry, you should use SLC format.
GRD data has been multi-looked and projected to ground range using an Earth ellipsoid model.
Pixel values represent the intensity but phase information is lost. You can use GRD data if you are interested in using only the intensity values (e.g perform classification), on the other hand, you can use the phase information if you are interested in doing a subsidence analysis.


but in this case what is the "Save as complex output option in the SLC to GRD tool for ?


SLC to GRD is just a graph so it’s made up of the operators that make up that graph. Therefore, those operators will still have the same input parameters they always have. In the case of the SLC to GRD graph, it will not make sense to output complex from the multilook operator.

Kind attn: bsaulquin, johngan,

orb>>CAL>>deburst>>(ML &/or any other speckle correction )>>TC has helped me retain phase information.

I am planning to use the above for H-α decomposition. The preprocessed image however contains scalloping effect (horizontal banding patterns or stripes). Have you guys faced this?

puneeth89 at gmail dot com

maybe you can try another input DEM for the terrain correction (SRTM Arc 1 or 3) to see if the scalloping effect persists.

1 Like

ABraun, thanks!! changed from 1 to arc 3 & the banding pattern vanished!

The horizontal banding pattern reappeared after “matrix generation” for a subset (subset_0_of_S1A_IW_SLC__ after _Orb_Cal_deb_ML_TC_mat) . I assume that the Matrix is just for calculations & not to be considered for visualization of any sort. Any thoughts on this!


It is either an acquisition artefact of this scene or it was somehow introduced during the graph. If you need exactly this scene, I recommend to compute the steps manually and check every intermediate result to see where it arises, Multi-Looking maybe.

Terrain correct as your last step. Also, try the matrix before the mulitlook. If you’re then going to do a pol decomposition - do it before the TC

I found that multi-looking causes stripes in some Sentinel-1 SLC data: Crop classification using SENTINEL-1A SLC data

We need to check/update the implementation of multilooking, it’s from the early days of NEST and notice that it has has a disclaimer saying that for complex products it is done “without resampling”.