DurLAR: A High-fidelity 128-channel LiDAR Dataset with Panoramic Ambient and Reflectivity Imagery for Multi-modal Autonomous Driving Applications

TL;DR

Evaluation shows the joint use supervised and self-supervised loss terms, enabled via the superior ground truth resolution and availability within DurLAR improves the quantitative and qualitative performance of leading contemporary monocular depth estimation approaches.

Abstract

We present DurLAR, a high-fidelity 128-channel 3D LiDAR dataset with panoramic ambient (near infrared) and reflectivity imagery, as well as a sample benchmark task using depth estimation for autonomous driving applications. Our driving platform is equipped with a high resolution 128 channel LiDAR, a 2MPix stereo camera, a lux meter and a GNSS/INS system. Ambient and reflectivity images are made available along with the LiDAR point clouds to facilitate multi-modal use of concurrent ambient and reflectivity scene information. Leveraging DurLAR, with a resolution exceeding that of prior benchmarks, we consider the task of monocular depth estimation and use this increased availability of higher resolution, yet sparse ground truth scene depth information to propose a novel joint supervised/self-supervised loss formulation. We compare performance over both our new DurLAR dataset, the established KITTI benchmark and the Cityscapes dataset. Our evaluation shows our joint use supervised and self-supervised loss terms, enabled via the superior ground truth resolution and availability within DurLAR improves the quantitative and qualitative performance of leading contemporary monocular depth estimation approaches (RMSE=3.639, Sq Rel=0.936).

Figures (14)

• Figure 1: LiDAR point clouds from two exemplar scenes with differing vertical LiDAR resolution (top to bottom: colour RGB images, [32 $\xrightarrow{}$ 64 $\xrightarrow{}$ 128] LiDAR channels).
• Figure 2: Test vehicle (Renault Twizy): equipped with a long range stereo camera, a LiDAR, a lux meter and a combined GNSS/INS inertial navigation system.
• Figure 3: Sensor placements, top view. All coordinate axes follow the right-hand rule (sizes in mm).
• Figure 4: The route (blue curves) used for dataset collection showing a variety of driving environments.
• Figure 5: Examples from DurLAR which demonstrate the diversity in our dataset. From top to bottom, RGB left camera images (top), grayscale right camera images (centre) and LiDAR point cloud (bottom). The point cloud is projected onto the 2D image plane using the LiDAR-to-left-camera external calibration, and the colour varies with the distance from the LiDAR (near$:=$red $\to$ far$:=$green).