HOTFLoc++: End-to-End Hierarchical LiDAR Place Recognition, Re-Ranking, and 6-DoF Metric Localisation in Forests

2025-11-12

AI Summary

To address the challenge of robust LiDAR-based localization in cluttered, self-similar, and cross-view environments—such as forests and urban scenes—this paper proposes an end-to-end framework comprising place recognition, re-ranking, and 6-degree-of-freedom (6-DoF) registration. Its key contributions are: (1) multi-granularity local descriptor extraction via an octree-based Transformer, enhancing robustness against occlusion and viewpoint variation; (2) a learnable multi-scale geometric verification module that mitigates re-ranking failure caused by degeneracy in single-scale matching; and (3) a coarse-to-fine registration strategy enabling ground-to-aerial cross-view localization. Evaluated on the CS-Wild-Places dataset, the method achieves a Recall@1 of 90.7% (+29.6% over baseline), with 97.2% of 6-DoF registrations achieving ≤2 m translation and ≤5° rotation error; re-ranking reduces average localization error by approximately half.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Computer Vision: Multi-modal Vision
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Systems and Infrastructure for Web, Mobile and WoT: Location- and context-aware Web and WoT applications and services
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

This article presents HOTFLoc++, an end-to-end framework for LiDAR place recognition, re-ranking, and 6-DoF metric localisation in forests. Leveraging an octree-based transformer, our approach extracts hierarchical local descriptors at multiple granularities to increase robustness to clutter, self-similarity, and viewpoint changes in challenging scenarios, including ground-to-ground and ground-to-aerial in forest and urban environments. We propose a learnable multi-scale geometric verification module to reduce re-ranking failures in the presence of degraded single-scale correspondences. Our coarse-to-fine registration approach achieves comparable or lower localisation errors to baselines, with runtime improvements of two orders of magnitude over RANSAC for dense point clouds. Experimental results on public datasets show the superiority of our approach compared to state-of-the-art methods, achieving an average Recall@1 of 90.7% on CS-Wild-Places: an improvement of 29.6 percentage points over baselines, while maintaining high performance on single-source benchmarks with an average Recall@1 of 91.7% and 96.0% on Wild-Places and MulRan, respectively. Our method achieves under 2 m and 5 degrees error for 97.2% of 6-DoF registration attempts, with our multi-scale re-ranking module reducing localisation errors by ~2$ imes$ on average. The code will be available upon acceptance.

Problem

• Achieving robust LiDAR place recognition in challenging forest environments with clutter
• Reducing localization errors through multi-scale geometric verification and re-ranking
• Enabling efficient 6-DoF metric localization with coarse-to-fine registration approach

Innovation

• Octree-based transformer extracts hierarchical local descriptors
• Learnable multi-scale geometric verification reduces re-ranking failures
• Coarse-to-fine registration achieves faster localization than RANSAC