CARLA Drone: Monocular 3D Object Detection from a Different Perspective

TL;DR

This paper introduces the CARLA Drone dataset to evaluate monocular 3D detection across diverse camera perspectives, and proposes GroundMix, a data augmentation method that improves detection accuracy, demonstrating the importance of perspective diversity.

Contribution

The paper presents the new CDrone dataset for drone-view 3D detection and a ground-based augmentation pipeline called GroundMix, advancing evaluation and training methods.

Findings

CDrone dataset challenges existing detection methods.

GroundMix significantly improves detection accuracy.

Models perform poorly across diverse perspectives without augmentation.

Abstract

Existing techniques for monocular 3D detection have a serious restriction. They tend to perform well only on a limited set of benchmarks, faring well either on ego-centric car views or on traffic camera views, but rarely on both. To encourage progress, this work advocates for an extended evaluation of 3D detection frameworks across different camera perspectives. We make two key contributions. First, we introduce the CARLA Drone dataset, CDrone. Simulating drone views, it substantially expands the diversity of camera perspectives in existing benchmarks. Despite its synthetic nature, CDrone represents a real-world challenge. To show this, we confirm that previous techniques struggle to perform well both on CDrone and a real-world 3D drone dataset. Second, we develop an effective data augmentation pipeline called GroundMix. Its distinguishing element is the use of the ground for creating 3D-consistent augmentation of a training image. GroundMix significantly boosts the detection accuracy of a lightweight one-stage detector. In our expanded evaluation, we achieve the average precision on par with or substantially higher than the previous state of the art across all tested datasets.