PLUMENet: Efficient 3D Object Detection from Stereo Images

3D object detection is a key component of many robotic applications such as self-driving vehicles. While many approaches rely on expensive 3D sensors such as LiDAR to produce accurate 3D estimates, methods that exploit stereo cameras have recently shown promising results at a lower cost. Existing approaches tackle this problem in two steps: first depth estimation from stereo images is performed to produce a pseudo LiDAR point cloud, which is then used as input to a 3D object detector. However, this approach is suboptimal due to the representation mismatch, as the two tasks are optimized in two different metric spaces. In this paper we propose a model that unifies these two tasks and performs them in the same metric space. Specifically, we directly construct a pseudo LiDAR feature volume (PLUME) in 3D space, which is then used to solve both depth estimation and object detection tasks. Our approach achieves state-of-the-art performance with much faster inference times when compared to existing methods on the challenging KITTI benchmark [1].

I. Introduction

Self-driving vehicles have the potential to revolutionize the future of mobility. In order to build a robust and safe autonomy system, reliable real-time perception is a necessity. Many self-driving cars are equipped with expensive LiDAR sensors, which provide accurate depth measurements that facilitate reasoning in 3D metric space. In recent years LiDAR based 3D detectors [2]–[6] have shown very promising results. However, LiDAR sensors remain costly and have limited range, limiting their applicability.