In the field of precision agriculture, weed detection models combined with selective spraying by ground or aerial robotics are seen as promising approaches for increasing yield harvests while simultaneously minimizing herbicide use. The available labeled training data are a major bottleneck for developing and applying supervised deep learning models, and more automated methods for labeled data generation are therefore needed. Our research aims to address this need by introducing a rule-based method for label data generation for perennial weeds. For this research, a dataset of a barley field was collected using an unmanned aerial vehicle (UAV) with a flight altitude of 10 m. A multispectral and a thermal camera were used for the data collection. The preprocessed dataset consists of multispectral and thermal orthomosaic images along with a canopy height model. The methodological part of this article introduces a proposed rule-based method for labeled data generation for perennial weeds based on the Normalized Difference Vegetation Index (NDVI), and this approach is further used to generate labels for the measured data. The preprocessed data combined with the generated labels was used to train U-net models. Three data combinations are used for the training and testing: multispectral, multispectral–thermal and multispectral–thermal–canopy–height–model. This approach was used to evaluate whether additional data improve model performance. To evaluate the models on ground truth labels, they are tested on a manually annotated test dataset, which consists of 10% of the whole dataset. The tested models achieved an F1 score of 0.82–0.83 on the test dataset. This finding demonstrates that the rule-based labeling method generates valid labels for the perennial weed detection task. Furthermore, our study shows that data fusion improved the model slightly. The data combination of multispectral–thermal–canopy–height–model as input resulted in the best-performing model, with an F1 score of 0.835.
The design of science-based policies to improve the sustainability of smallholder agriculture is challenged by a limited understanding of fundamental system properties, such as the spatial distribution of active cropland and field size. We integrate 1.5 m resolution satellite imagery, satellite embeddings, and deep learning to map individual fields in complex smallholder systems at policy-relevant scale. We identify 17 million fields across Mozambique for the target year 2023, reflecting the cropland distribution but also pointing to previously uncaptured agricultural regions located in agricultural frontier regions which host 5%–9% of the Mozambican population. Field size in Mozambique is low overall, with half of the fields being smaller than 0.2 ha, and 78% smaller than 0.5 ha. Field size varies with accessibility, population density, and forest cover change, suggesting the presence of diverse actors, including semi-subsistence smallholder farms, medium-scale commercial farms, and large-scale farming operations. Our results suggest that field size is a key indicator relating to the socio-economic and environmental outcomes of agriculture and their trade-offs.