CAFTS – A COHERENCE AND FEATURE TRACKING STUDY FOR SENTINEL-1

摘要

Temporal decorrelation is a major problem when applying repeat pass interferometry or coherent feature tracking techniques to ice and snow surfaces as it significantly reduces the amount of viable data. This study was performed to assess the level of coherence measured between SAR image pairs with a 6 and 12-day temporal baseline, in line with the planned repeat cycles of Sentinel-1 (S-1). Based on this, an assessment was made about the proportion of S-1 data that will be suitable for measuring ice velocity. Analysis showed high levels of coherence over the ‘exposed rock’ land cover type, with 75.5% of data from the 12-day image pairs remaining above the 0.7 coherence threshold. High levels of coherence were measured from 3 and 6-day image pairs over the ‘ice sheet’ and ‘ice stream’ land surface types however, over 12-days just 42.5% and 19.1% of data remained above the 0.7 coherence threshold respectively. This highlights the importance of utilising the S-1 6-day repeat period for obtaining sufficiently coherent data over snow and ice land cover types.