Joint Monocular 3D Vehicle Detection and Tracking

TL;DR

This paper introduces an online monocular video framework for 3D vehicle detection and tracking, utilizing 3D box matching, trajectory prediction, and motion learning to improve accuracy and robustness in dynamic environments.

Contribution

It presents a novel monocular 3D vehicle detection and tracking method with 3D matching, trajectory re-identification, and LSTM-based motion prediction, outperforming LiDAR-based methods within 30 meters.

Findings

Robust data association and tracking in diverse datasets.

Significant improvement over LiDAR-based methods within 30 meters.

Effective long-term motion extrapolation using LSTM.

Abstract

Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods.