Real-Time Multi-Modal Semantic Fusion on Unmanned Aerial Vehicles with Label Propagation for Cross-Domain Adaptation

TL;DR

This paper presents a real-time UAV system that fuses multi-modal sensor data for semantic scene understanding, utilizing label propagation for cross-domain adaptation, achieving high-frequency semantic mapping in complex environments.

Contribution

The work introduces a lightweight, real-time multi-modal semantic fusion system on UAVs with a novel label propagation method for cross-domain sensor adaptation.

Findings

Achieves approximately 9 Hz semantic inference and fusion onboard UAVs.

Demonstrates effective semantic mapping in urban and disaster environments.

Validates system performance through real-world experiments.

Abstract

Unmanned aerial vehicles (UAVs) equipped with multiple complementary sensors have tremendous potential for fast autonomous or remote-controlled semantic scene analysis, e.g., for disaster examination. Here, we propose a UAV system for real-time semantic inference and fusion of multiple sensor modalities. Semantic segmentation of LiDAR scans and RGB images, as well as object detection on RGB and thermal images, run online onboard the UAV computer using lightweight CNN architectures and embedded inference accelerators. We follow a late fusion approach where semantic information from multiple sensor modalities augments 3D point clouds and image segmentation masks while also generating an allocentric semantic map. Label propagation on the semantic map allows for sensor-specific adaptation with cross-modality and cross-domain supervision. Our system provides augmented semantic images and point clouds with $\approx$ 9 Hz. We evaluate the integrated system in real-world experiments in an urban environment and at a disaster test site.