LodoNet: A Deep Neural Network with 2D Keypoint Matchingfor 3D LiDAR Odometry Estimation

TL;DR

LodoNet introduces a novel approach for LiDAR odometry by converting 3D point clouds into 2D images, extracting keypoints with SIFT, and using these for accurate odometry estimation, achieving competitive results on KITTI.

Contribution

The paper presents a new method that reformulates LiDAR odometry as an image feature extraction problem using 2D keypoints, improving accuracy and efficiency.

Findings

Achieves state-of-the-art or comparable results on KITTI benchmark.

Utilizes SIFT for precise 2D keypoint matching in LiDAR data.

Demonstrates the effectiveness of image-based feature extraction for 3D odometry.

Abstract

Deep learning based LiDAR odometry (LO) estimation attracts increasing research interests in the field of autonomous driving and robotics. Existing works feed consecutive LiDAR frames into neural networks as point clouds and match pairs in the learned feature space. In contrast, motivated by the success of image based feature extractors, we propose to transfer the LiDAR frames to image space and reformulate the problem as image feature extraction. With the help of scale-invariant feature transform (SIFT) for feature extraction, we are able to generate matched keypoint pairs (MKPs) that can be precisely returned to the 3D space. A convolutional neural network pipeline is designed for LiDAR odometry estimation by extracted MKPs. The proposed scheme, namely LodoNet, is then evaluated in the KITTI odometry estimation benchmark, achieving on par with or even better results than the state-of-the-art.