Geographical Context Matters: Bridging Fine and Coarse Spatial Information to Enhance Continental Land Cover Mapping

TL;DR

BRIDGE-LC tackles continental-scale LULC mapping by integrating geospatial metadata into a lightweight MLP framework. It introduces a positional encoding module, a region-specific branch, and a region-invariant land cover classifier trained with a supervised contrastive loss to disentangle location-dependent features. Evaluations on EU-27 LUCAS 2022 data in EXTRAP and LORO settings show that combining fine-grained coordinates with coarse-grained biogeographical regions yields the best accuracy, especially for crop-type classification, while maintaining computational efficiency. The findings demonstrate that multi-scale geospatial context substantially improves generalization and detail in large-scale LULC products, offering a practical approach for scalable, region-aware land cover mapping.

Abstract

Land use and land cover mapping from Earth Observation (EO) data is a critical tool for sustainable land and resource management. While advanced machine learning and deep learning algorithms excel at analyzing EO imagery data, they often overlook crucial geospatial metadata information that could enhance scalability and accuracy across regional, continental, and global scales. To address this limitation, we propose BRIDGE-LC (Bi-level Representation Integration for Disentangled GEospatial Land Cover), a novel deep learning framework that integrates multi-scale geospatial information into the land cover classification process. By simultaneously leveraging fine-grained (latitude/longitude) and coarse-grained (biogeographical region) spatial information, our lightweight multi-layer perceptron architecture learns from both during training but only requires fine-grained information for inference, allowing it to disentangle region-specific from region-agnostic land cover features while maintaining computational efficiency. To assess the quality of our framework, we use an open-access in-situ dataset and adopt several competing classification approaches commonly considered for large-scale land cover mapping. We evaluated all approaches through two scenarios: an extrapolation scenario in which training data encompasses samples from all biogeographical regions, and a leave-one-region-out scenario where one region is excluded from training. We also explore the spatial representation learned by our model, highlighting a connection between its internal manifold and the geographical information used during training. Our results demonstrate that integrating geospatial information improves land cover mapping performance, with the most substantial gains achieved by jointly leveraging both fine- and coarse-grained spatial information.

Figures (11)

• Figure 1: Database samples, distributed over the EU-27 countries.
• Figure 2: Subset of the biogeographical regions produced by the European Environment Agency (EEA).
• Figure 3: Proposed framework. In addition to the standard land cover classification branch (yellow), which processes reflectance-based input features, we introduce two complementary components: (1) a fine-grained spatial information branch (green) utilizing latitude-longitude coordinates, and (2) a coarse-grained spatial information branch (red) based on biogeographical region labels. During training, the red branch guides the model to learn region-invariant embeddings that are more robust and generalizable for land cover classification. This branch is used only during training and can be discarded at inference time.
• Figure 4: Confusion matrices (row-normalized) for Level-1 classification, extrapolation scenario.
• Figure 5: Confusion matrices (row-normalized) for Level-2 classification, extrapolation scenario.