AV-DTEC: Self-Supervised Audio-Visual Fusion for Drone Trajectory Estimation and Classification

TL;DR

AV-DTEC introduces a lightweight self-supervised audio-visual fusion system for drone detection, enhancing robustness and accuracy in real-world conditions by integrating multi-modal features and adaptive weighting.

Contribution

It presents a novel self-supervised learning framework with a plug-and-play feature enhancement module and a teacher-student model for improved drone detection.

Findings

High accuracy in real-world multi-modality data

Effective cross-lighting robustness

Open-source code and models available

Abstract

The increasing use of compact UAVs has created significant threats to public safety, while traditional drone detection systems are often bulky and costly. To address these challenges, we propose AV-DTEC, a lightweight self-supervised audio-visual fusion-based anti-UAV system. AV-DTEC is trained using self-supervised learning with labels generated by LiDAR, and it simultaneously learns audio and visual features through a parallel selective state-space model. With the learned features, a specially designed plug-and-play primary-auxiliary feature enhancement module integrates visual features into audio features for better robustness in cross-lighting conditions. To reduce reliance on auxiliary features and align modalities, we propose a teacher-student model that adaptively adjusts the weighting of visual features. AV-DTEC demonstrates exceptional accuracy and effectiveness in real-world multi-modality data. The code and trained models are publicly accessible on GitHub \url{https://github.com/AmazingDay1/AV-DETC}.