DisDop: Distillation with Domain Priors for Open-Vocabulary Aerial Object Detection

2026-05-23

AI Summary

Open-vocabulary aerial object detection remains challenging due to the scarcity of aerial imagery and its significant distributional shift from natural images, rendering generic approaches ineffective. This work proposes DisDop, a unified framework that, for the first time, jointly leverages RemoteCLIP and DINOv3 to integrate visual, textual, and global contextual information through multi-level domain-prior knowledge distillation. The method introduces a teacher fusion strategy, models textual semantic relationships, and incorporates a global context enhancement mechanism to achieve cross-modal alignment and fine-grained feature extraction. Evaluated on open-vocabulary aerial object detection benchmarks, DisDop establishes a new state of the art, with ablation studies confirming the effectiveness and novelty of its individual components.

Abstract

With the widespread application of drones in recent years, object detection of aerial images has attracted increasing attention, especially open-vocabulary aerial detection which is not restricted to predefined categories. Due to the scarcity of drone's viewpoint images and their significant differences from natural images, it is difficult to achieve satisfying results by directly applying vanilla open-vocabulary detection methods designed for natural scenarios. Some studies propose to transfer knowledge from pre-trained models by using lightweight networks or generating pseudo labels, but they tend to rely on models trained on natural images, neglecting the potential of foundation models specifically tailored for remote sensing and aerial imagery. To address this limitation, we propose DisDop, a unified framework that systematically distills multi-level domain priors from remote sensing foundation models (e.g., RemoteCLIP and DINOv3) into a lightweight detector. Specifically, we first distill visual priors through a teacher fusion strategy that combines RemoteCLIP's cross-modal alignment capability with DINOv3's fine-grained local feature extraction ability, transferring their complementary strengths to the detector's backbone. Second, we distill textual priors embedded in RemoteCLIP's text encoder by explicitly modeling inter-category semantic relationships, while incorporating global contextual priors to enhance local feature representation for small objects. Through this multi-level prior distillation framework, our DisDop achieves new state-of-the-art performance on open-vocabulary aerial detection benchmarks. Extensive ablation analysis also demonstrates the rationality and effectiveness of our proposed modules.

Problem

• open-vocabulary detection
• aerial object detection
• domain priors
• foundation models
• knowledge distillation

Innovation

• open-vocabulary detection
• knowledge distillation
• domain priors
• aerial object detection
• foundation models