High-Res Satellite Data Slashes Farm Boundary Errors
A new dataset proves that 3m resolution imagery drastically improves the mapping of smallholder farms across 24 countries.
Higher resolution satellite imagery significantly improves the accuracy of field boundary mapping, a critical component for crop monitoring and yield estimation. The Fields of the Planet (FTP) dataset utilizes 3m PlanetScope imagery to resolve smallholder parcels that are often invisible or blurred in standard 10m Sentinel-2 pixels.
https://arxiv.org/abs/2607.04449
Switching to 3m imagery raises panoptic quality (PQ) from 21.0 to 35.5 and reduces matched-boundary error from 18.6m to 7.4m. The gains are most pronounced for the smallest plots; for fields under 0.5 hectares, PQ jumps from 5.8 to 15.7. This precision allows for more reliable vectorization of land use across 133,168 targets.
These technical gains directly impact the precision of agricultural commodities intelligence. Accurate delineation of small-scale farming is essential for irrigation planning and global supply forecasts, particularly in regions where fragmented land ownership makes traditional mapping unreliable.