Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Using Multi-Temporal Sentinel-1 and Sentinel-2 data for water bodies mapping

Luigi Russo, Francesco Mauro, Babak Memar, Alessandro Sebastianelli, Paolo Gamba, Silvia Liberata Ullo

TL;DR

The paper tackles robust water body mapping under climate-change-driven variability, where clouds and weather challenge optical sensing. It proposes a multisensor multitemporal dataset by fusing Sentinel-1 SAR with Sentinel-2 multispectral imagery, extending the SEN2DWATER framework over six years to create a 329×39×300×300×15 data cube. Benchmarking uses the SWI and NDWI indices and an unsupervised k-means classifier on the combined data, validated against the dynamicworld GT dataset, demonstrating high accuracy and the relative strength of NDWI. The work highlights the value of integrated SAR-optical data for improved water resource monitoring and sets the stage for global expansion and DL-based enhancements.

Abstract

Climate change is intensifying extreme weather events, causing both water scarcity and severe rainfall unpredictability, and posing threats to sustainable development, biodiversity, and access to water and sanitation. This paper aims to provide valuable insights for comprehensive water resource monitoring under diverse meteorological conditions. An extension of the SEN2DWATER dataset is proposed to enhance its capabilities for water basin segmentation. Through the integration of temporally and spatially aligned radar information from Sentinel-1 data with the existing multispectral Sentinel-2 data, a novel multisource and multitemporal dataset is generated. Benchmarking the enhanced dataset involves the application of indices such as the Soil Water Index (SWI) and Normalized Difference Water Index (NDWI), along with an unsupervised Machine Learning (ML) classifier (k-means clustering). Promising results are obtained and potential future developments and applications arising from this research are also explored.

Using Multi-Temporal Sentinel-1 and Sentinel-2 data for water bodies mapping

TL;DR

The paper tackles robust water body mapping under climate-change-driven variability, where clouds and weather challenge optical sensing. It proposes a multisensor multitemporal dataset by fusing Sentinel-1 SAR with Sentinel-2 multispectral imagery, extending the SEN2DWATER framework over six years to create a 329×39×300×300×15 data cube. Benchmarking uses the SWI and NDWI indices and an unsupervised k-means classifier on the combined data, validated against the dynamicworld GT dataset, demonstrating high accuracy and the relative strength of NDWI. The work highlights the value of integrated SAR-optical data for improved water resource monitoring and sets the stage for global expansion and DL-based enhancements.

Abstract

Climate change is intensifying extreme weather events, causing both water scarcity and severe rainfall unpredictability, and posing threats to sustainable development, biodiversity, and access to water and sanitation. This paper aims to provide valuable insights for comprehensive water resource monitoring under diverse meteorological conditions. An extension of the SEN2DWATER dataset is proposed to enhance its capabilities for water basin segmentation. Through the integration of temporally and spatially aligned radar information from Sentinel-1 data with the existing multispectral Sentinel-2 data, a novel multisource and multitemporal dataset is generated. Benchmarking the enhanced dataset involves the application of indices such as the Soil Water Index (SWI) and Normalized Difference Water Index (NDWI), along with an unsupervised Machine Learning (ML) classifier (k-means clustering). Promising results are obtained and potential future developments and applications arising from this research are also explored.
Paper Structure (4 sections, 2 equations, 5 figures, 2 tables)

This paper contains 4 sections, 2 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Dataset Creation
  3. Dataset Benchmarking
  4. CONCLUSIONS

Figures (5)

  • Figure 1: Visualization of the new dataset. Different geographical locations are represented on the y-axis, while the instants of each time series are represented on the x-axis for each location. Each cube plot shows the Sen1 & Sen2 compound product composed of 13 spectral + 2 polarization bands.
  • Figure 2: Our Sentinel-1 and Sentinel-2 datacube.
  • Figure 3: Workflow of the first and second methods based on the computation of the SWI and NDWI indices.
  • Figure 4: Workflow of the third method based on k-means clustering algorithm.
  • Figure 5: Visual results for the three proposed methods in comparison with the GT.