SLCDeepETC: An On-Demand Analysis Ready Data Pipeline on Sentinel-1 Single Look Complex for Deep Learning

SLCDeepETC is an on-demand Analysis Ready Data pipeline designed to automate data processing and cross-platform delivery of interferometry data on Single Look Complex products from Sentinel-1 to predict some environmental phenomena using Deep Learning. By retaining all original sensor measurements, it has been proven that interferometry data on Single Look Complex products, when analyzed with Deep Learning, can better inform data restoration, coherence estimation, classification, and automatic target recognition. However, the large data volume, interferometry processing complexity, interferometry data interpretation difficulties, and heterogeneous framework codebases for Deep Learning algorithms pose challenges to Deep Learning model training, hindering radar interferometry domain research. Through the ETC (Extract, Transform, and Cross-platform delivery) pipeline mechanism, SLCDeepETC preserves essential details and achieves significant speedups and superior performance for radar interferometry analysis with Deep Learning on real-world raw Sentinel-1 Single Look Complex data products. The proposed SLCDeepETC pipeline, on demand, can generate large Analysis Ready Datasets of interferometry derivative time series data on Single Look Complex products for Deep Learning model training through inter-framework codebase functionality, accelerating Deep Learning research in the radar interferometry domain.