Collaborative Engineering Project

Context

The NASA Harvest program aims to analyze geospatial data in order to contribute to global food security. In a team of four engineering students, the project focused on agricultural crops in Ukraine, using satellite imagery in the context of the ongoing armed conflict.

Goals

Quantitatively evaluate an algorithm that automatically delineates agricultural boundaries from satellite images.

Approach

Research, study, and implementation of geospatial data comparison metrics in Python. Creation of reference geospatial datasets via QGIS and Planetscope in order to compare them with the outputs generated by the algorithm under evaluation. Oral presentations, exchanges with the supervising team, and technical report writing.

Results

Algorithm accuracy determined per metric and per Ukrainian oblast (approx. 90%). Featured collaboration on the NASA Harvest website Co-author in Nature