Towards Robust Deep Learning-based Rumex Obtusifolius Detection from Drone Images

2026-04-28

AI Summary

This study addresses the performance degradation of weed recognition models—trained on ground-vehicle-collected data—when applied to drone imagery due to domain shift. The authors investigate domain adaptation strategies to enhance cross-domain robustness and find that Vision Transformers pretrained with self-supervision (DINOv2/DINOv3) inherently mitigate domain shift, significantly outperforming conventional CNNs combined with explicit domain adaptation techniques such as moment matching or maximum classifier discrepancy. Experimental results on the authors’ newly introduced drone-based dataset, AGSMultiRumex, demonstrate that this approach achieves an F1 score of 0.8. To foster further research in this area, the AGSMultiRumex dataset has been made publicly available.

Abstract

Domain adaptation (DA) addresses the challenge of transferring a machine learning model trained on a source domain to a target domain with a different data distribution. In this work, we study DA for the task of Rumex obtusifolius (Rumex) image classification. We train models on a published, ground vehicle-based dataset (source) and evaluate their performance on a custom target dataset acquired by unmanned aerial vehicles (UAVs). We find that Convolutional Neural Network (CNN) models, specifically ResNets, generalize poorly to the target domain, even after fine-tuning on the source data. Applying moment-matching and maximum classifier discrepancy, two established DA techniques, substantially improves target-domain performance. However, Vision Transformer (ViT) models pretrained with self-supervised objectives (DINOv2, DINOv3) handle domain shifts intrinsically well, surpassing even moment-matching-trained ResNets, likely due to the rich, general-purpose representations acquired during large-scale pretraining. Using ViTs fine-tuned on the source dataset, we demonstrate high classification performances in the range of F1=0.8 on our target dataset. To support further research on DA for weed detection in grassland systems, we publicly release our UAV-based target dataset AGSMultiRumex, comprising data from 15 flights over Swiss meadows.

Problem

• domain adaptation
• Rumex obtusifolius
• drone imagery
• domain shift
• weed detection

Innovation

• Domain Adaptation
• Vision Transformer
• Self-supervised Pretraining
• Rumex Detection
• UAV Imagery