D3DLO: Deep 3D LiDAR Odometry
This work addresses LiDAR odometry for autonomous vehicles or robotics, but it is incremental as it builds on existing deep learning approaches with efficiency improvements.
The authors tackled the problem of LiDAR odometry by proposing a deep learning architecture that directly processes 3D point clouds, achieving similar performance to a prior method while using only about 3.56% of the network parameters and reducing input size by up to 50% with a marginal performance decrease.
LiDAR odometry (LO) describes the task of finding an alignment of subsequent LiDAR point clouds. This alignment can be used to estimate the motion of the platform where the LiDAR sensor is mounted on. Currently, on the well-known KITTI Vision Benchmark Suite state-of-the-art algorithms are non-learning approaches. We propose a network architecture that learns LO by directly processing 3D point clouds. It is trained on the KITTI dataset in an end-to-end manner without the necessity of pre-defining corresponding pairs of points. An evaluation on the KITTI Vision Benchmark Suite shows similar performance to a previously published work, DeepCLR [1], even though our model uses only around 3.56% of the number of network parameters thereof. Furthermore, a plane point extraction is applied which leads to a marginal performance decrease while simultaneously reducing the input size by up to 50%.