ICRA 2019 论文速览 | 传统SLAM、三维视觉算法进展

点击上方“计算机视觉life”，选择“星标”

快速获得最新干货

笔者汇总了ICRA 2019 SLAM相关论文，总共分为四个部分：

Deep learning + traditional SLAM，见 ICRA 2019 论文速览 | SLAM 爱上 Deep Learning

Deep learning based SLAM

Traditional SLAM/3D vision

SLAM evaluation and datasets

本文介绍:Traditional SLAM

1. SLAM——Direct, 2D/3D feature, Lidar SLAM

FMD Stereo SLAM: Fusing MVG and Direct Formulation towards Accurate and Fast Stereo SLAM（中科院，特征点法和直接法结合）

Keywords: SLAM, Localization, Mapping

RESLAM: A Real-Time Robust Edge-Based SLAM System （边缘SLAM）

Keywords: SLAM, Visual-Based Navigation, RGB-D Perception

代码：
https://github.com/fabianschenk/RESLAM
https://github.com/fabianschenk/REVO

Leveraging Structural Regularity of Atlanta World for Monocular SLAM（Atlanta世界坐标系下的边缘线约束SLAM）

Keywords: SLAM, Localization, Mapping

Illumination Robust Monocular Direct Visual Odometry for Outdoor Environment Mapping（抗光照变化的直接法视觉里程计）

Loosely-Coupled Semi-Direct Monocular SLAM——IRAL（直接法跟踪，特征点法做地图优化和回环）

Keywords: SLAM, Localization, Mapping

代码
https://github.com/sunghoon031/LCSD_SLAM

3D Keypoint Repeatability for Heterogeneous Multi-Robot SLAM（多机器人系统的不同传感器的特征点匹配，3D关键点KPQ-SI和NARF两个特征点比较适合用于Loopclosure和多机器人重定位中）

Keywords: SLAM, Performance Evaluation and Benchmarking, Mapping

Local Descriptor for Robust Place Recognition Using LiDAR Intensity——IRAL （ISHOT点云描述子用于定位）

Keywords: Localization, Field Robots, Autonomous Vehicle Navigation

1-Day Learning, 1-Year Localization: Long-Term LiDAR Localization Using Scan Context Image——IRAL（激光雷达的长期定位方法）

Keywords: Localization, Range Sensing, SLAM

2. SLAM——Pose Optimization

On-Line 3D Active Pose-Graph SLAM Based on Key Poses Using Graph Topology and Sub-Maps（位姿优化，子地图）

Keywords: SLAM, Motion and Path Planning

MH-iSAM2: Multi-Hypothesis iSAM Using Bayes Tree and Hypo-Tree（非线性增量优化，解决SLAM歧义）

Keywords: SLAM, Localization, Mapping

Visual SLAM: Why Bundle Adjust?（BA的替代优化方法，解决纯旋转和弱平移下的位姿估计）

Keywords: SLAM

Modeling Perceptual Aliasing in SLAM Via Discrete-Continuous Graphical Models——IRAL （离散连续图模型的优化方法）

Keywords: SLAM, Sensor Fusion, Optimization and Optimal Control

POSEAMM: A Unified Framework for Solving Pose Problems Using an Alternating Minimization Method（使用交替最小化方法解决姿势优化问题的统一框架）

Keywords: Computer Vision for Automation, Omnidirectional Vision, Localization

Visual-Odometric Localization and Mapping for Ground Vehicles Using SE(2)-XYZ Constraints（平面移动机器人的位姿估计约束模型）

Keywords: Localization, SLAM, Sensor Fusion

https://github.com/izhengfan/se2lam

Direct Relative Edge Optimization, a Robust Alternative for Pose Graph Optimization（边缘约束的图优化）

Keywords: SLAM, Mapping, Multi-Robot Systems

A White-Noise-On-Jerk Motion Prior for Continuous-Time Trajectory Estimation on SE(3) （位姿估计方法）

Keywords: SLAM

Low-Latency Visual SLAM with Appearance-Enhanced Local Map Building（一种快速局部地图的策略）

Keywords: SLAM

3. SLAM——VIO/ VISLAM

Fast and Robust Initialization for Visual-Inertial SLAM（VINS初始化）

Keywords: SLAM, Mapping, Localization

Visual-Inertial Navigation: A Concise Review

Keywords: Autonomous Vehicle Navigation, Localization, Sensor Fusion
https://github.com/rpng

Tightly-Coupled Aided Inertial Navigation with Point and Plane Features（点面特征的VINS系统）

Keywords: Range Sensing, Sensor Fusion, SLAM

Tightly-Coupled Visual-Inertial Localization and 3D Rigid-Body Target Tracking——IRAL（VINS和跟踪物体紧融合）

Keywords: Localization, Visual Tracking, SLAM

Aided Inertial Navigation: Unified Feature Representations and Observability Analysis（点，线，面多特征融合的VINS系统）

Keywords: Localization, SLAM, Visual-Based Navigation

A Linear-Complexity EKF for Visual-Inertial Navigation with Loop Closures（一种MSCKF的VINS方法，带回环）

Keywords: Localization, SLAM, Mapping

Multi-Camera Visual-Inertial Navigation with Online Intrinsic and Extrinsic Calibration（多相机VINS系统的在线标定相机，IMU内外参数方法）

Keywords: Visual-Based Navigation, Sensor Fusion, Localization

Sensor-Failure-Resilient Multi-IMU Visual-Inertial Navigation（一种多IMU的VINS系统）

Keywords: Localization, SLAM, Failure Detection and Recovery

Efficient 2D-3D Matching for Multi-Camera Visual Localization（多camera imu的重定位）

Keywords: Localization, Computer Vision for Transportation, Omnidirectional Vision

Keyframe-Based Direct Thermal–Inertial Odometry（低质量图像下的VIO方法，基于关键帧的直接法，可以借鉴他借鉴低照度下的vo问题）

Keywords: Localization, Sensor Fusion, Field Robots

Improving the Robustness of Visual-Inertial Extended Kalman Filtering（VINS 系统姿态估计提升方案）

Keywords: Visual-Based Navigation, Aerial Systems: Perception and Autonomy, Robust/Adaptive Control of Robotic Systems

Towards Fully Dense Direct Filter-Based Monocular Visual-Inertial Odometry（密集直接法VINS系统）

Keywords: Sensor Fusion, Visual-Based Navigation, Localization

Experimental Comparison of Visual-Aided Odometry Methods for Rail Vehicles—IRAL  （在火车的数据集上实验比对VO、VIO方法）

Keywords: Computer Vision for Transportation, Intelligent Transportation Systems, SLAM

RaD-VIO: Rangefinder-Aided Downward Visual-Inertial Odometry（测距融合VIO）

Keywords: Aerial Systems: Perception and Autonomy, Localization, Performance Evaluation and Benchmarking

4. SLAM——Multi-sensor Fusion

Accurate Direct Visual-Laser Odometry with Explicit Occlusion Handling and Plane Detection（激光雷达融合视觉定位，区分平面和非平面的特征点）

Keywords: SLAM, Localization

Robust Pose-Graph SLAM Using Absolute Orientation Sensing（激光雷达+天花板摄像头SLAM）

Keywords: SLAM, Industrial Robots

Tightly Coupled 3D Lidar Inertial Odometry and Mapping（雷达和IMU融合）

Keywords: Computer Vision for Automation, Sensor Fusion, SLAM

IN2LAMA: INertial Lidar Localisation and Mapping（IMU和Lidar融合的SLAM）

Keywords: Mapping, SLAM, Sensor Fusion

ROVO: Robust Omnidirectional Visual Odometry for Wide-Baseline Wide-FOV Camera Systems（多鱼眼SLAM）

Keywords: SLAM, Omnidirectional Vision, Autonomous Vehicle Navigation

5. Depth & Mapping & 3D Reconstruction

ScalableFusion: High-Resolution Mesh-Based Real-Time 3D Reconstruction（三维重建）

Keywords: SLAM, Mapping, RGB-D Perception

Surfel-Based Dense RGB-D Reconstruction with Global and Local Consistency（用SFM计算全局的关键帧位姿，同时用slam方法计算局部相邻帧的位姿，然后用FGO factor graph optimization方法将全局和局部信息融合计算出密集三维重建）

Keywords: SLAM, Localization, Mapping

Real-Time Scalable Dense Surfel Mapping

Keywords: Mapping, Sensor Fusion, Aerial Systems: Perception and Autonomy

代码
https://github.com/HKUST-Aerial-Robotics/DenseSurfelMapping

Real-Time Dense Mapping for Self-Driving Vehicles Using Fisheye Cameras（鱼眼相机的密集三维重建）

Keywords: Mapping, Computer Vision for Transportation, Omnidirectional Vision

https://zhpcui.github.io/projects/arxiv18_densemapping/

Real Time Dense Depth Estimation by Fusing Stereo with Sparse Depth Measurements（用TOF辅助双目密集匹配算法）

Keywords: Range Sensing, Aerial Systems: Perception and Autonomy

Dense Surface Reconstruction from Monocular Vision and LiDAR（雷达和视觉融合三维重建）

Keywords: Mapping, SLAM, Range Sensing

Incremental Visual-Inertial 3D Mesh Generation with Structural Regularities（VIO输出的稀疏点做三维重建的三角面片）

Keywords: SLAM, Visual-Based Navigation, Sensor Fusion

OVPC Mesh: 3D Free-Space Representation for Local Ground Vehicle Navigation（3D Mesh表示方法，用于无人车避障）

Keywords: Autonomous Vehicle Navigation, Field Robots, Mapping

KO-Fusion: Dense Visual SLAM with Tightly-Coupled Kinematic and Odometric Tracking（机器人运动学与里程计结合的密集三维重建）

Keywords: SLAM, Sensor Fusion, Kinematics

3D Surface Reconstruction Using a Two-Step Stereo Matching Method Assisted with Five Projected Patterns（三维双目结构光重建设备）

Keywords: Computer Vision for Automation, Range Sensing, Computer Vision for Other Robotic Applications

6. Localization——Lidar / Vision

Beyond Point Clouds: Fisher Information Field for Active Visual Localization（3D landmark来做视觉定位）

Keywords: Visual-Based Navigation, Localization, Motion and Path Planning

Effective Visual Place Recognition Using Multi-Sequence Maps—IRAL（场景识别定位）

Keywords: Localization

MRS-VPR: A Multi-Resolution Sampling Based Visual Place Recognition Method（场景识别和回环检测，高效、多尺度、粗到细的长期序列VPR）

Keywords: SLAM, Deep Learning in Robotics and Automation, Visual Learning

Probabilistic Appearance-Based Place Recognition through Bag of Tracked Words——IRAL （BTW场景定位）

Keywords: SLAM, Visual-Based Navigation, Recognition

Geometric Relation Distribution for Place Recognition——IRAL（激光雷达的重定位和回环）

Keywords: Mapping, Localization, Range Sensing

代码
https://github.com/dlr1516/grd

7. Others

A-SLAM: Human-In-The-Loop Augmented SLAM（交互式SLAM地图和位姿修正方法）

Keywords: SLAM, Virtual Reality and Interfaces, Wheeled Robots

Iteratively Reweighted Midpoint Method for Fast Multiple View Triangulation——IRAL （三角化误差消除方法）

Keywords: SLAM, Mapping

CELLO-3D: Estimating the Covariance of ICP in the Real World（点云ICP）

Keywords: SLAM, Range Sensing, Learning and Adaptive Systems

交流群

欢迎加入公众号读者群一起和同行交流，目前有SLAM、算法竞赛、图像检测分割、人脸人体、医学影像、自动驾驶、综合等微信群（以后会逐渐细分），请扫描下面微信号加群，备注：”昵称+学校/公司+研究方向“，例如：”张三 + 上海交大 + 视觉SLAM“。请按照格式备注，否则不予通过。添加成功后会根据研究方向邀请进入相关微信群。请勿在群内发送广告，否则会请出群，谢谢理解~

从零开始学习三维视觉核心技术SLAM，扫描查看介绍，3天内无条件退款

学习切忌单打独斗，一个良好的学习圈子能够帮助你快速入门，交流讨论才能少走弯路，快速进步！

如有AI领域实习、求职、招聘、项目合作、咨询服务等需求，快来加入我们吧，期待和你建立连接，找人找技术不再难！

推荐阅读

吐血整理 | SLAM方向国内有哪些优秀的公司？

最强战队 | 三维视觉、SLAM方向全球顶尖实验室汇总

从零开始一起学习SLAM | 为什么要学SLAM？

从零开始一起学习SLAM | 学习SLAM到底需要学什么？

从零开始一起学习SLAM | SLAM有什么用？

从零开始一起学习SLAM | C++新特性要不要学？

从零开始一起学习SLAM | 为什么要用齐次坐标？

从零开始一起学习SLAM | 三维空间刚体的旋转

从零开始一起学习SLAM | 为啥需要李群与李代数？

从零开始一起学习SLAM | 相机成像模型

从零开始一起学习SLAM | 不推公式，如何真正理解对极约束?

从零开始一起学习SLAM | 神奇的单应矩阵

从零开始一起学习SLAM | 你好，点云

从零开始一起学习SLAM | 给点云加个滤网

从零开始一起学习SLAM | 点云平滑法线估计

从零开始一起学习SLAM | 点云到网格的进化

从零开始一起学习SLAM | 理解图优化，一步步带你看懂g2o代码

从零开始一起学习SLAM | 掌握g2o顶点编程套路

从零开始一起学习SLAM | 掌握g2o边的代码套路

从零开始一起学习SLAM | ICP原理及应用

                                                                        最新AI干货，我在看