GeoUniPR：几何一致的跨模态地点识别统一框架
原标题：GeoUniPR: A Geometry-Consistent Unified Framework for Cross-Modal Place Recognition
AI 摘要
GeoUniPR 提出了一种几何一致的跨模态地点识别框架，通过将 LiDAR 点云投影到相机视角构建多通道深度图像视图，并利用参数高效微调的 ViT 编码器学习统一嵌入空间，无需复杂的对齐模块或多阶段训练。实验表明，GeoUniPR 在 KITTI 和 KITTI-360 数据集上达到了最先进的性能，并具有良好的跨数据集泛化能力。
正文节选
GeoUniPR: A Geometry-Consistent Unified Framework for Cross-Modal Place Recognition Abstract Cross-modal place recognition (CMPR) aims to identify the same location across heterogeneous sensing modalities, such as vision and LiDAR. Existing methods commonly bridge the modality gap using complex alignment modules, multi-stage training, or full fine-tuning of pretrained backbones. In this work, we revisit CMPR from the perspective of geometric consistency and propose GeoUniPR, a unified and concis