Uncertainty-aware Bayesian machine learning modelling of land cover classification
This work addresses the need for traceability in metrology for land cover mapping, offering a more trustworthy model for remote sensing applications, though it is incremental as it adapts an existing Bayesian method to a specific domain.
The authors tackled the problem of land cover classification by proposing a Bayesian framework that incorporates input measurement uncertainty, which is often ignored by existing models. They applied Bayesian quadratic discriminant analysis to Sentinel-2 datasets and found that it maintains predictive performance across years and sizes while being more interpretable and computationally efficient.
Land cover classification involves the production of land cover maps, which determine the type of land through remote sensing imagery. Over recent years, such classification is being performed by machine learning classification models, which can give highly accurate predictions on land cover per pixel using large quantities of input training data. However, such models do not currently take account of input measurement uncertainty, which is vital for traceability in metrology. In this work we propose a Bayesian classification framework using generative modelling to take account of input measurement uncertainty. We take the specific case of Bayesian quadratic discriminant analysis, and apply it to land cover datasets from Copernicus Sentinel-2 in 2020 and 2021. We benchmark the performance of the model against more popular classification models used in land cover maps such as random forests and neural networks. We find that such Bayesian models are more trustworthy, in the sense that they are more interpretable, explicitly model the input measurement uncertainty, and maintain predictive performance of class probability outputs across datasets of different years and sizes, whilst also being computationally efficient.