Dune Reclaimer

Problem

Desertification monitoring involves understanding how landscapes change across both space and time.

Objective

Develop a deep learning pipeline to investigate desertification patterns using satellite imagery.

Approach

Prepare spatiotemporal datasets from Sentinel-2 imagery and use ConvLSTM to model environmental change.

Implementation

Processed satellite imagery, prepared temporal inputs, and developed a ConvLSTM-based pipeline for desertification prediction.

Challenges to investigate

Temporal consistency and comparable imagery are important considerations when preparing satellite-image sequences. The CV does not describe how these issues were resolved.

The source does not document the project’s actual obstacles; the following is a technical consideration, not a retrospective claim.

Documented outcomes

The CV documents satellite-image processing and development of the deep learning pipeline. No verified accuracy, deployment, or field-impact figures were provided.

Technical reflection

Spatiotemporal models need data preparation that preserves both the spatial structure and the ordering of observations.

This reflection follows from the documented method; it is not presented as a measured project outcome.