Remote Sensing

This project compares and integrates different remote sensing sources to estimate soil properties under varying conditions. We evaluate multispectral optical data, hyperspectral imagery, and radar signals to understand which features best predict targets such as moisture, SOC, salinity indicators, and surface condition. Field measurements are linked to sensor-derived features (bands, indices, time series, bare-soil composites) and modelled using machine learning with spatially robust validation. Outputs include sensor-specific and fused soil maps plus practical guidance on when each sensor works best.

Main focus

• Use multi-sensor remote sensing to estimate soil properties at scale.
• Compare sensors and fusion strategies under realistic validation.
• Deliver actionable soil maps and sensor-selection guidance.

Objectives

1. Build comparable feature sets from optical, hyperspectral, and radar data.
2. Train ML models to evaluate sensor performance and data fusion benefits.
3. Produce soil maps (and confidence) that scale from fields to regions.

Graphical abstract

:::: {.row .justify-content-sm-center} ::: {.col-sm-10 .mt-3 .mt-md-0}

::: ::::

::: caption Graphical abstract summarizing the workflow and key results. :::

Papers (selected)

Paper 1 — Multitemporal Sentinel-2 SWIR-Based SOC Estimation

DOI

Paper 2 — PRISMA Hyperspectral Mapping of Soil Organic Carbon

DOI

Paper 3 — Optimizing Rain Gauge Networks Using Remote Sensing

DOI