Biological Valuation Map of Flanders: A Sentinel-2 Imagery Analysis

TL;DR

The paper addresses the lack of dense, region-scale ground-truth data for land-use/land-cover (LULC) mapping by constructing the Biological Valuation Map (BVM) for Flanders and pairing it with Sentinel-2 imagery. It presents a reproducible data workflow, including polygon rasterization to 14 habitat classes, CRS harmonization, cloud masking, and a map-sheet based partitioning strategy for training, validation, and testing, using a U-Net with a VGG16 encoder across 3, 3-channel RG NIR, and 11-channel inputs. Preliminary results show pixel accuracies around 67–69%, with limited gains from additional spectral channels likely due to model capacity and fusion strategy, while highlighting class imbalance and compute as key challenges. The work provides a publicly available dataset and pipeline that enable reproducible regional remote-sensing ML research and benchmark comparisons across regions.

Abstract

In recent years, machine learning has become crucial in remote sensing analysis, particularly in the domain of Land-use/Land-cover (LULC). The synergy of machine learning and satellite imagery analysis has demonstrated significant productivity in this field, as evidenced by several studies. A notable challenge within this area is the semantic segmentation mapping of land usage over extensive territories, where the accessibility of accurate land-use data and the reliability of ground truth land-use labels pose significant difficulties. For example, providing a detailed and accurate pixel-wise labeled dataset of the Flanders region, a first-level administrative division of Belgium, can be particularly insightful. Yet there is a notable lack of regulated, formalized datasets and workflows for such studies in many regions globally. This paper introduces a comprehensive approach to addressing these gaps. We present a densely labeled ground truth map of Flanders paired with Sentinel-2 satellite imagery. Our methodology includes a formalized dataset division and sampling method, utilizing the topographic map layout 'Kaartbladversnijdingen,' and a detailed semantic segmentation model training pipeline. Preliminary benchmarking results are also provided to demonstrate the efficacy of our approach.

Figures (3)

• Figure 1: Example of a rasterized BVM map (different colors represent different pixel classes)
• Figure 2: A demonstration of map sheet cutouts (Source: Digital Flanders Agency)
• Figure 3: Loss and accuracy of 3 modes