RDC-UNet++: An end-to-end network for multispectral satellite image enhancement

• DOI: 10.1016/j.rsase.2024.101293
• Publication Date: 2024-07-09T16:18:15Z
• Authors: Shilpa Suresh; Ragesh Rajan M.; Asha C.S.; Fabio Dell’Acqua
• Keywords: Panchromatic film; Spectral bands

Multi-spectral satellite imagery is an ideal data source for comprehensive, real-time Earth observation (EO) due to its extensive coverage of and regular updates. It has a wide range applications in environment monitoring, disaster management, urban planning, weather forecasting etc. Yet, the visual aspect these images thus possibility extract useful information using image processing techniques often degraded fog, rain, dust, cloud, Satellite enhancement denotes set designed improve quality such that result more analysis. The aims enhanced analysis than original particular remote sensing application. This study introduces multi-spectral architecture called Residual Dense Connection-based UNet++ (RDC-UNet++). unique design can by enhancing their color texture details. Extensive experimental studies on datasets containing 150 prove proposed performs better recent state-of-the-art algorithms.