Search for: All records

Mapping agricultural fields using high-resolution satellite imagery and deep learning (DL) models has advanced significantly, even in regions with small, irregularly shaped fields. However, effective DL models often require large, expensive labeled datasets, which are typically limited to specific years or regions. This restricts the ability to create annual maps needed for agricultural monitoring, as changes in farming practices and environmental conditions cause domain shifts between years and locations. To address this, we focused on improving model generalization without relying on yearly labels through a holistic approach that integrates several techniques, including an area-based loss function, Tversky-focal loss (TFL), data augmentation, and the use of regularization techniques like dropout. Photometric augmentations helped encode invariance to brightness changes but also increased the incidence of false positives. The best results were achieved by combining photometric augmentation, TFL, and Monte Carlo dropout, although dropout alone led to more false negatives. Input normalization also played a key role, with the best results obtained when normalization statistics were calculated locally (per chip) across all bands. Our U-Net-based workflow successfully generated multi-year crop maps over large areas, outperforming the base model without photometric augmentation or MC-dropout by 17 IoU points. 

Food production data — such as crop, livestock, aquaculture and fisheries statistics — are critical to achieving multiple sustainable development goals. However, the lack of reliable, regularly collected, accessible, usable and spatially disaggregated statistics limits an accurate picture of the state of food production in many countries and prevents the implementation of effective food system interventions. In this Review, we take stock of national and international food production data to understand its availability and limitations. Across databases, there is substantial global variation in data timeliness, granularity (both spatially and by food category) and transparency. Data scarcity challenges are most pronounced for livestock and aquatic food production. These challenges are largely concentrated in Central America, the Middle East and Africa owing to a combination of inconsistent census implementation and a global reliance on self-reporting. Because data scarcity is the result of technical, institutional and political obstacles, solutions must include technological and policy innovations. Fusing traditional and emerging data-gathering techniques with coordinated governance and dedicated long-term financing will be key to overcoming current obstacles to sustained, up-to-date and accurate food production data collection, foundational in promoting and monitoring progress towards healthier and more sustainable food systems worldwide.