Problem
Research questions and friction points this paper is trying to address.
LiDAR Odometry and Mapping
reconstruction fidelity
supervision signal
large-scale complex scenes
Hi-LOAM: Hierarchical Implicit Neural Fields for LiDAR Odometry and Mapping
🤖 AI Summary
This work proposes Hi-LOAM, a multi-scale implicit neural framework for LiDAR odometry and mapping that overcomes the limitations of existing methods, which often rely on supervised signals or suffer from insufficient reconstruction fidelity in complex large-scale environments. Hi-LOAM takes raw LiDAR point clouds as input and learns hierarchical implicit features stored in multi-level hash tables organized via an octree structure, which are then decoded into signed distance fields using a shallow MLP. Pose estimation is achieved through correspondence-free scan-to-implicit alignment, and the entire system is trained end-to-end in a self-supervised manner without explicit data association or external supervision. Evaluated on multiple real-world and synthetic datasets, Hi-LOAM significantly outperforms state-of-the-art approaches, demonstrating superior reconstruction accuracy and strong cross-scene generalization capabilities.
📝 Abstract
LiDAR Odometry and Mapping (LOAM) is a pivotal technique for embodied-AI applications such as autonomous driving and robot navigation. Most existing LOAM frameworks are either contingent on the supervision signal, or lack of the reconstruction fidelity, which are deficient in depicting details of large-scale complex scenes. To overcome these limitations, we propose a multi-scale implicit neural localization and mapping framework using LiDAR sensor, called Hi-LOAM. Hi-LOAM receives LiDAR point cloud as the input data modality, learns and stores hierarchical latent features in multiple levels of hash tables based on an octree structure, then these multi-scale latent features are decoded into signed distance value through shallow Multilayer Perceptrons (MLPs) in the mapping procedure. For pose estimation procedure, we rely on a correspondence-free, scan-to-implicit matching paradigm to estimate optimal pose and register current scan into the submap. The entire training process is conducted in a self-supervised manner, which waives the model pre-training and manifests its generalizability when applied to diverse environments. Extensive experiments on multiple real-world and synthetic datasets demonstrate the superior performance, in terms of the effectiveness and generalization capabilities, of our Hi-LOAM compared to existing state-of-the-art methods.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Implicit Neural Fields
LiDAR Odometry and Mapping
Self-supervised Learning
Hierarchical Hash Encoding
Scan-to-Implicit Registration
🔎 Similar Papers
No similar papers found.
