TimePillars: Temporally-Recurrent 3D LiDAR Object Detection

TL;DR

TimePillars introduces a temporally-recurrent 3D LiDAR object detection method that enhances long-range detection by aggregating multiple scans while maintaining efficiency, leveraging the Zenseact Open Dataset.

Contribution

The paper presents a novel recurrent detection pipeline using pillar representation that improves long-range detection in LiDAR data efficiently.

Findings

Recurrency improves detection robustness.

Basic building blocks suffice for effective performance.

Method achieves real-time long-range detection.

Abstract

Object detection applied to LiDAR point clouds is a relevant task in robotics, and particularly in autonomous driving. Single frame methods, predominant in the field, exploit information from individual sensor scans. Recent approaches achieve good performance, at relatively low inference time. Nevertheless, given the inherent high sparsity of LiDAR data, these methods struggle in long-range detection (e.g. 200m) which we deem to be critical in achieving safe automation. Aggregating multiple scans not only leads to a denser point cloud representation, but it also brings time-awareness to the system, and provides information about how the environment is changing. Solutions of this kind, however, are often highly problem-specific, demand careful data processing, and tend not to fulfil runtime requirements. In this context we propose TimePillars, a temporally-recurrent object detection pipeline which leverages the pillar representation of LiDAR data across time, respecting hardware integration efficiency constraints, and exploiting the diversity and long-range information of the novel Zenseact Open Dataset (ZOD). Through experimentation, we prove the benefits of having recurrency, and show how basic building blocks are enough to achieve robust and efficient results.