发表机构
SK Intellix（SK智能科技）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对3D点云在有限或不对称视场下全局定位难的问题，提出G-PROBE框架。通过虚拟传感器分解、跨视场分支集合等方法，结合确定性耦合定位，无需学习。在多数据集和模式上评估，该框架性能出色，端到端可用性高，在不对称视场下优势明显。
AI 中文摘要
在有限或不对称视场（FOV）下，从3D点云进行全局定位仍然具有挑战性，因为这些视场无法提供位置识别方法所假设的密集、对称覆盖。我们提出了G-PROBE，一个无需学习的全局定位框架，消除了这一假设。虚拟传感器分解通过设计在从窄视场传感器到全景或多传感器装置的配置上运行相同的管道。前端枚举跨视场分支集合，为航向不变的位置识别编码航向假设。一个分数尺度不变、无需调整的gamma-SGRT在部分视场下抑制航向混叠，并在对称360度时可证明变为惰性。后端CG-GICP通过限制在由鸟瞰确定性图（前端评分的副产品）选择的高确定性共同观测点上的一次传递来细化粗略的全云GICP。这种确定性耦合将描述符评估与6自由度度量姿态估计联系起来，而无需外部验证模块。在五个LiDAR数据集和三种模式（机械、固态、FMCW）上进行评估，G-PROBE平均获得最高的无需学习的多会话F1，并且在全景单会话设置中具有竞争力。在手工制作和零样本监督基线在宽到窄的跨传感器配对下崩溃的情况下，它仍然可以端到端使用（成功率高达55.0%，而最强的无需学习的基线不超过6.8%），并且在视场不对称（360度到60度）下，它保留约54%的Recall@1，约为最强的无需学习的基线的18倍。
英文摘要
Under field-of-view (FOV) mismatch, pooling LiDAR features over unequal angular support can distort compact retrieval keys and exclude correct matches before geometric verification. We present PROBE-X, a learning-free method for single-scan cross-FOV place recognition. Building on PROBE's probabilistic occupancy representation, which models translation uncertainty, PROBE-X constructs ring-mean retrieval keys using angular masks conditioned on up to four fixed hypotheses for the unknown relative heading. Candidate lists are merged by maximum cosine similarity. Each retained candidate is aligned under the hypotheses that retrieved it and assigned a geometric score over the angular overlap at the refined heading. Across controlled-FOV and heterogeneous-LiDAR experiments, PROBE-X achieves higher Recall@1 and PR-AUC than the evaluated single-scan baselines in most settings with limited-FOV queries and panoramic database scans. On the controlled asymmetric-FOV sweep, single-scan PROBE-X also outperforms the baselines evaluated with up to ten accumulated scans on both metrics. In two ablation settings, removing the angular masks from retrieval-key construction reduces candidate coverage and Recall@1, even when alignment and scoring still account for angular support. Project page: https://sites.google.com/view/probex-pr
CommentsSubstantially revised and retitled. Focused on single-scan cross-FOV place recognition, with updated retrieval design and evaluation. The localization backend of v1 is not included. 8 pages, 5 figures, 5 tables
