UniLGL: Learning Uniform Place Recognition for FOV-limited/Panoramic LiDAR Global Localization

2025-07-16

AI Summary

Existing LiDAR-based global localization (LGL) methods rely heavily on geometric features or assume homogeneous sensors, failing to simultaneously ensure consistency across spatial layout, material properties, and sensor modalities. Method: We propose UniLGL—the first LGL framework to replace translation equivariance with viewpoint invariance, enabling unified descriptor learning across heterogeneous LiDARs (e.g., narrow-FOV vs. panoramic). It introduces a multi-BEV fusion network that encodes point clouds into spatial-intensity dual-channel bird’s-eye-view (BEV) images for end-to-end robust feature extraction, and a registration-free SE(3) pose estimator that jointly leverages 2D–3D feature mapping to solve global pose. Contribution/Results: UniLGL achieves state-of-the-art performance on real-world port and forest benchmarks. It has been successfully deployed on large-scale trucks and micro aerial vehicles, enabling high-precision localization, mapping, and multi-robot collaborative exploration.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision
• Machine Learning: Multimodal Learning

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Existing LGL methods typically consider only partial information (e.g., geometric features) from LiDAR observations or are designed for homogeneous LiDAR sensors, overlooking the uniformity in LGL. In this work, a uniform LGL method is proposed, termed UniLGL, which simultaneously achieves spatial and material uniformity, as well as sensor-type uniformity. The key idea of the proposed method is to encode the complete point cloud, which contains both geometric and material information, into a pair of BEV images (i.e., a spatial BEV image and an intensity BEV image). An end-to-end multi-BEV fusion network is designed to extract uniform features, equipping UniLGL with spatial and material uniformity. To ensure robust LGL across heterogeneous LiDAR sensors, a viewpoint invariance hypothesis is introduced, which replaces the conventional translation equivariance assumption commonly used in existing LPR networks and supervises UniLGL to achieve sensor-type uniformity in both global descriptors and local feature representations. Finally, based on the mapping between local features on the 2D BEV image and the point cloud, a robust global pose estimator is derived that determines the global minimum of the global pose on SE(3) without requiring additional registration. To validate the effectiveness of the proposed uniform LGL, extensive benchmarks are conducted in real-world environments, and the results show that the proposed UniLGL is demonstratively competitive compared to other State-of-the-Art LGL methods. Furthermore, UniLGL has been deployed on diverse platforms, including full-size trucks and agile Micro Aerial Vehicles (MAVs), to enable high-precision localization and mapping as well as multi-MAV collaborative exploration in port and forest environments, demonstrating the applicability of UniLGL in industrial and field scenarios.

Problem

• Achieves uniform LiDAR global localization across sensor types
• Encodes complete point cloud into BEV images for feature extraction
• Ensures robust pose estimation without additional registration steps

Innovation

• Encodes point cloud into spatial and intensity BEV images
• Uses multi-BEV fusion for uniform feature extraction
• Introduces viewpoint invariance for sensor-type uniformity