A general framework for map-based visual localization. It contains 1) Map Generation which support traditional features or deeplearning features. 2) Hierarchical-Localizationvisual in visual(points or line) map. 3)Fusion framework with IMU, wheel odom and GPS sensors.

A simple system that can relocalize a robot on a built map is developed in this system. The system is based on LIO-SAM.

[ECCV 2022] Map-free Visual Relocalization: Metric Pose Relative to a Single Image

MixVPR: Feature Mixing for Visual Place Recognition (WACV 2023)

Official repository for the ICLR 2024 paper "Towards Seamless Adaptation of Pre-trained Models for Visual Place Recognition".

Official repository for the CVPR 2024 paper "CricaVPR: Cross-image Correlation-aware Representation Learning for Visual Place Recognition".

DH3D: Deep Hierarchical 3D Descriptors for Robust Large-Scale 6DOF Relocalization

深度学习和三维视觉相关的论文

The framework for visual place recognition in changing enviroments. Matches two sequence of images of arbitrary trajectory overlap.

[ICRA 2025 & RAL 2024] D2S: Representing sparse descriptors and 3D coordinates for camera relocalization

[TRO] Fast and Accurate Deep Loop Closing and Relocalization for Reliable LiDAR SLAM

[JAG 2023] LiDAR localization at 100FPS: A map-aided and template descriptor-based global method

Official repository for the AAAI 2024 paper "Deep Homography Estimation for Visual Place Recognition".

Code for the CoRL 2021 paper "SeqMatchNet: Contrastive Learning with Sequence Matching for Place Recognition \& Relocalization"

[3DV21] Visual Camera Re-Localization Using Graph Neural Networks and Relative Pose Supervision, M. Türkoǧlu et al.

Official repository for the NeurIPS 2024 paper "SuperVLAD: Compact and Robust Image Descriptors for Visual Place Recognition".

Homography-based loss function for camera pose regression

This is a robotic package for an algorithm for visual teach and repeat

This a simple implementation of V-LOAM

Visual-Lidar Map Alignment