DROW: Real-Time Deep Learning based Wheelchair Detection in 2D Range Data

TL;DR

The paper presents DROW, a deep learning-based detector for 2D laser range data that achieves state-of-the-art results in wheelchair and walker detection, with broad applicability to other classes.

Contribution

It introduces a CNN-based detection method with novel preprocessing and voting schemes for 2D range data, surpassing traditional handcrafted feature approaches.

Findings

Achieved state-of-the-art detection accuracy for wheelchairs and walkers.

Demonstrated effectiveness of CNNs in laser range data detection tasks.

Provided a large annotated dataset and ROS implementation for the community.

Abstract

We introduce the DROW detector, a deep learning based detector for 2D range data. Laser scanners are lighting invariant, provide accurate range data, and typically cover a large field of view, making them interesting sensors for robotics applications. So far, research on detection in laser range data has been dominated by hand-crafted features and boosted classifiers, potentially losing performance due to suboptimal design choices. We propose a Convolutional Neural Network (CNN) based detector for this task. We show how to effectively apply CNNs for detection in 2D range data, and propose a depth preprocessing step and voting scheme that significantly improve CNN performance. We demonstrate our approach on wheelchairs and walkers, obtaining state of the art detection results. Apart from the training data, none of our design choices limits the detector to these two classes, though. We provide a ROS node for our detector and release our dataset containing 464k laser scans, out of which 24k were annotated.