Task-Relevant and Irrelevant Region-Aware Augmentation for Generalizable Vision-Based Imitation Learning in Agricultural Manipulation

TL;DR

This paper introduces DRAIL, a region-aware augmentation framework that improves the generalization of vision-based imitation learning in agricultural robotics by focusing on task-relevant features and reducing background bias.

Contribution

DRAIL explicitly separates and augments task-relevant and irrelevant regions to enhance policy robustness in agricultural manipulation tasks.

Findings

DRAIL improves success rates under unseen visual conditions.

Policies rely more on task-essential features, increasing robustness.

Enhanced generalization demonstrated in real robot experiments.

Abstract

Vision-based imitation learning has shown promise for robotic manipulation; however, its generalization remains limited in practical agricultural tasks. This limitation stems from scarce demonstration data and substantial visual domain gaps caused by i) crop-specific appearance diversity and ii) background variations. To address this limitation, we propose Dual-Region Augmentation for Imitation Learning (DRAIL), a region-aware augmentation framework designed for generalizable vision-based imitation learning in agricultural manipulation. DRAIL explicitly separates visual observations into task-relevant and task-irrelevant regions. The task-relevant region is augmented in a domain-knowledge-driven manner to preserve essential visual characteristics, while the task-irrelevant region is aggressively randomized to suppress spurious background correlations. By jointly handling both sources of visual variation, DRAIL promotes learning policies that rely on task-essential features rather than incidental visual cues. We evaluate DRAIL on diffusion policy-based visuomotor controllers through robot experiments on artificial vegetable harvesting and real lettuce defective leaf picking preparation tasks. The results show consistent improvements in success rates under unseen visual conditions compared to baseline methods. Further attention analysis and representation generalization metrics indicate that the learned policies rely more on task-essential visual features, resulting in enhanced robustness and generalization.