Kuva Space takes on illicit crop monitoring with hyperspectral satellites and AI

Finnish Kuva Space completed a pilot with a European customer: Hyperfield-1 hyperspectral images and AI classification identified poppy fields in Afghanistan's Helmand. The system annotated 248,889 plots and marked 8,835 as probable poppy, with accuracy rising to 75% versus 71% for Sentinel-2.

• 363 hyperspectral images taken January–June 2026, covering 648,450 km²
• 248,889 fields annotated, 8,835 marked as probable opium poppy
• Accuracy 75% versus 71% using only Sentinel-2
• Goal is at least 90% accuracy with Hyperfield-2 satellites and SAR data