AerialFusionMapNet: Online HD Map Construction with Aerial-Onboard BEV Fusion

2026-06-23

AI Summary

Existing end-to-end fusion approaches fail to fully exploit structural priors inherent in aerial imagery, limiting their performance in online high-definition map construction. This work proposes a structured two-stage training framework that explicitly enhances the contribution of aerial features to bird’s-eye-view (BEV) representations within a unified end-to-end pipeline, enabling effective fusion between aerial and vehicle-mounted sensor BEV features. By optimizing the training strategy rather than increasing model complexity, the method significantly improves the utilization efficiency of aerial priors. On the nuScenes geographically partitioned benchmark, the model achieves 54.7 mAP, yielding an absolute improvement of 5.9 mAP (a relative gain of 12.1%) over current fusion baselines.

Abstract

High-resolution aerial imagery has recently emerged as a complementary modality for automated driving perception and has shown potential to improve birds-eye-view (BEV) scene understanding when fused with onboard sensors. Prior work demonstrated performance gains for online high-definition (HD) map construction through aerial-onboard fusion; however, conventional end-to-end fusion does not fully exploit the structural information contained in aerial representations. In this work, we introduce AerialFusionMapNet, a fusion-based mapping framework with a structured two-stage training strategy that explicitly enhances the contribution of aerial features within a unified pipeline. The proposed training scheme enables more effective integration of structural aerial priors. On the nuScenes geographic split, AerialFusionMapNet achieves up to 54.7 mAP, improving over prior aerial-onboard fusion baselines from 48.8 mAP by +5.9 absolute and +12.1% relative. The results suggest that structured training design, rather than increased architectural complexity, plays a more decisive role in unlocking the full potential of aerial imagery for online HD map construction. Code and trained models are available at https://github.com/DriverlessMobility/AerialFusionMapNet.

Problem

• HD map construction
• aerial-onboard fusion
• BEV scene understanding
• aerial imagery
• online mapping

Innovation

• Aerial-onboard fusion
• structured training
• HD map construction
• BEV perception
• two-stage training