Edits made because I noticed that on uncorrected IFGs, the unw.pngs were very different between original and loopy IFG.
Thought this was just a plotting issue, but it was a result of the multi-looking used (I was using an unw_thresh of 0.1, vs 0.5 in LiCSBAS). Have changed this now so that the same ML thresholds are used.
Side benefit could be that a low threshold was allowing more bad pixels through (good pixels generally seemed to be occurring in a high enough densty anyway, that although more were being allowed through, the increase wasn't to the same amount) , and this could mean that fewer IFGs are selected. Increasing to be in line with LiCSBAS should mean that more corrected IFGs are accepted
Edits made because I noticed that on uncorrected IFGs, the unw.pngs were very different between original and loopy IFG. Thought this was just a plotting issue, but it was a result of the multi-looking used (I was using an unw_thresh of 0.1, vs 0.5 in LiCSBAS). Have changed this now so that the same ML thresholds are used.
Side benefit could be that a low threshold was allowing more bad pixels through (good pixels generally seemed to be occurring in a high enough densty anyway, that although more were being allowed through, the increase wasn't to the same amount) , and this could mean that fewer IFGs are selected. Increasing to be in line with LiCSBAS should mean that more corrected IFGs are accepted