LRPD: Long Range 3D Pedestrian Detection Leveraging Specific Strengths of LiDAR and RGB

TL;DR

This paper introduces LRPD, a novel method for long-range 3D pedestrian detection in autonomous vehicles, combining RGB and LiDAR data to improve accuracy at extended distances, evaluated on the KITTI benchmark.

Contribution

The paper presents a new approach that leverages RGB and LiDAR strengths for enhanced long-range pedestrian detection, outperforming existing methods.

Findings

Significant improvement in mAP on long-range pedestrians

Effective fusion of RGB segmentation and LiDAR proposals

Validated on KITTI benchmark data

Abstract

While short range 3D pedestrian detection is sufficient for emergency breaking, long range detections are required for smooth breaking and gaining trust in autonomous vehicles. The current state-of-the-art on the KITTI benchmark performs suboptimal in detecting the position of pedestrians at long range. Thus, we propose an approach specifically targeting long range 3D pedestrian detection (LRPD), leveraging the density of RGB and the precision of LiDAR. Therefore, for proposals, RGB instance segmentation and LiDAR point based proposal generation are combined, followed by a second stage using both sensor modalities symmetrically. This leads to a significant improvement in mAP on long range compared to the current state-of-the art. The evaluation of our LRPD approach was done on the pedestrians from the KITTI benchmark.