Hedgementation = Hedgerow Segmentation: A Remote Sensing Benchmark

2026-06-22

AI Summary

This study addresses the absence of a national-scale, high-resolution benchmark for hedgerow segmentation in remote sensing—a fine-grained agricultural land cover feature—hindering the evaluation of model generalization across diverse geographic and climatic conditions. To bridge this gap, we present the first nationwide hedgerow segmentation benchmark for France at a 10 m² spatial resolution, integrating multi-source remote sensing data with harmonized ground survey labels to support both supervised and self-supervised learning paradigms. Leveraging this benchmark, we systematically evaluate the cross-regional and cross-climatic transfer performance of three baseline models. The code and results are publicly released to establish a new standard and provide a robust tool for high-precision remote sensing monitoring of agroecological features.

Abstract

We propose Hedgementation: a new benchmark to evaluate machine learning models for hedgerow mapping from remote sensing data at country scale and 10m$^2$ spatial resolution. We combine and harmonize multiple remote sensing data products and ground truth labels sourced from a hedgerow inventory in France. We measure the ability of three baseline models to generalize across spatial distance, and across climatic zones, a more explicitly challenging task. Our benchmark tests both supervised and self-supervised learning approaches for remote sensing, applied to tracking fine-scale features of high agricultural importance. The code to reproduce the benchmark and baselines results is available at https://github.com/hedgementation/hedgementation.

Problem

• hedgerow segmentation
• remote sensing
• spatial generalization
• climatic zones
• fine-scale features

Innovation

• hedgerow segmentation
• remote sensing benchmark
• self-supervised learning
• spatial generalization
• high-resolution mapping