Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Mapping Farmed Landscapes from Remote Sensing

Michelangelo Conserva, Alex Wilson, Charlotte Stanton, Vishal Batchu, Varun Gulshan

TL;DR

This paper tackles the lack of accessible, fine-scale maps of farmland landscapes necessary for biodiversity targets. It introduces Farmscapes, the England-wide 25 cm resolution map of hedgerows, stone walls, woodlands, and farmed land, produced with a segmentation model trained on 942 annotated tiles and grounded with LiDAR reference data. The approach achieves high F1-scores for woodland (96) and farmed land (95), with more moderate scores for hedgerows (72) and stone walls (60), and is released openly via Google Earth Engine for ecologists and policymakers. The work enables data-driven habitat restoration planning and sets the stage for landscape connectivity analyses, while outlining limitations and directions for geographic expansion and change-monitoring.

Abstract

Effective management of agricultural landscapes is critical for meeting global biodiversity targets, but efforts are hampered by the absence of detailed, large-scale ecological maps. To address this, we introduce Farmscapes, the first large-scale (covering most of England), high-resolution (25cm) map of rural landscape features, including ecologically vital elements like hedgerows, woodlands, and stone walls. This map was generated using a deep learning segmentation model trained on a novel, dataset of 942 manually annotated tiles derived from aerial imagery. Our model accurately identifies key habitats, achieving high f1-scores for woodland (96\%) and farmed land (95\%), and demonstrates strong capability in segmenting linear features, with an F1-score of 72\% for hedgerows. By releasing the England-wide map on Google Earth Engine, we provide a powerful, open-access tool for ecologists and policymakers. This work enables data-driven planning for habitat restoration, supports the monitoring of initiatives like the EU Biodiversity Strategy, and lays the foundation for advanced analysis of landscape connectivity.

Mapping Farmed Landscapes from Remote Sensing

TL;DR

This paper tackles the lack of accessible, fine-scale maps of farmland landscapes necessary for biodiversity targets. It introduces Farmscapes, the England-wide 25 cm resolution map of hedgerows, stone walls, woodlands, and farmed land, produced with a segmentation model trained on 942 annotated tiles and grounded with LiDAR reference data. The approach achieves high F1-scores for woodland (96) and farmed land (95), with more moderate scores for hedgerows (72) and stone walls (60), and is released openly via Google Earth Engine for ecologists and policymakers. The work enables data-driven habitat restoration planning and sets the stage for landscape connectivity analyses, while outlining limitations and directions for geographic expansion and change-monitoring.

Abstract

Effective management of agricultural landscapes is critical for meeting global biodiversity targets, but efforts are hampered by the absence of detailed, large-scale ecological maps. To address this, we introduce Farmscapes, the first large-scale (covering most of England), high-resolution (25cm) map of rural landscape features, including ecologically vital elements like hedgerows, woodlands, and stone walls. This map was generated using a deep learning segmentation model trained on a novel, dataset of 942 manually annotated tiles derived from aerial imagery. Our model accurately identifies key habitats, achieving high f1-scores for woodland (96\%) and farmed land (95\%), and demonstrates strong capability in segmenting linear features, with an F1-score of 72\% for hedgerows. By releasing the England-wide map on Google Earth Engine, we provide a powerful, open-access tool for ecologists and policymakers. This work enables data-driven planning for habitat restoration, supports the monitoring of initiatives like the EU Biodiversity Strategy, and lays the foundation for advanced analysis of landscape connectivity.

Paper Structure

This paper contains 15 sections, 11 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Landscape elements
  3. Data
  4. Results
  5. Conclusion
  6. Annotation methodology and guidelines
  7. Related works
  8. Training details
  9. Hyperparameter optimization.
  10. Detailed performance metrics
  11. Ground Layer.
  12. Stone Walls.
  13. Hedgerows.
  14. Woodland.
  15. Qualitative examples

Figures (11)

  • Figure 1: Comparison of data sources.
  • Figure 2: Typical English landscape elements, hedgerows (left), stone walls (center), and woodland (right).
  • Figure 3: Types of hedgerows.
  • Figure 4: Sampling strategy illustration.
  • Figure 5: Model architecture.
  • ...and 6 more figures