xiaojake Goto Github PK

object-slam

!!基于RGB-D的人工环境SLAM:Compositional and Scalable Object SLAM (ICRA 2021)

obstacle_detector

!基于2D激光scan的动态障碍物检测:A ROS package for 2D obstacle detection based on laser range data.

occlusions-4d

显示遮挡Revealing Occlusions with 4D Neural Fields (CVPR 2022) - Official Implementation

occupancy_grid_mapping

!栅格地图构建&去除激光雷达畸变Here is the method to build a grid map and get rid of the laser distortion

odmd

!从相机运动和目标检测中获取深度:Object Depth via Motion and Detection Dataset

oh_my_loam

!去ros后的loam:ROS-free implementation of LOAM

omnimotion

识别并追踪任意物体

once_3dlanes_benchmark

3D车道线单目检测方法ONCE-3DLanes

onepose

!!无CAD模型的one-shot物体姿态估计Code for "OnePose: One-Shot Object Pose Estimation without CAD Models", CVPR 2022

open_vins

!基于MSCKF的vins方法:An open source platform for visual-inertial navigation research.

opencood

[ICRA 2022] An opensource framework for cooperative detection. Official implementation for OPV2V.

openpcdet

3D目标检测点云检测新网络PV-RCNN:OpenPCDet Toolbox for LiDAR-based 3D Object Detection.

openslam_parallaxba

ParallaxBA repos from OpenSLAM.org

opensource_slam_noted

!开源激光SLAM算法注释:open source slam system notes

openvslam

多平台适用的VSLAM:A Versatile Visual SLAM Framework

optimized-sc-f-loam

optimized_icp

!采用高斯牛顿的优化方法对icp进行了实现,其精度明显高于pcl库中基于svd分解的icp. 但计算资源的消耗和实时性上相对要差一些

orb_slam2_detailed_comments

!ORB-SLAM2 超详细注释:Detailed comments for ORB-SLAM2 with trouble-shooting, key formula derivation, and diagrammatic drawing

orbslamm

!!定位丢失后在线建图并融合到老地图的orbslam2

orepa_cvpr2022

!在线卷积重新参数化OREPA，节省70%的显存！训练速度提高2倍！CVPR 2022 "Online Convolutional Re-parameterization"

ouster_example

Ouster sample code

ov2slam

!!!(sota)在线实时多功能,单目,深度:OV²SLAM is a Fully Online and Versatile Visual SLAM for Real-Time Applications

overlap_localization

!!!OverlapNet,激光SLAM闭环检测,全局定位,AMXL:chen2020iros: Learning an Overlap-based Observation Model for 3D LiDAR Localization.

overlapnet

基于OverlapNet网络的3D激光回环检测算法:OverlapNet - Loop Closing for 3D LiDAR-based SLAM (chen2020rss)

overlaptransformer

!!基于网络的激光回环检测,具有旋转不变性The code for our paper submitted to RAL/IROS 2022: OverlapTransformer: An Efficient and Rotation-Invariant Transformer Network for LiDAR-Based Place Recognition.

pacc-net

Source code of "EdgeFormer: Improving Light-weight ConvNets by Learning from Vision Transformers"

paddledetection

!!!目标检测:Object detection and instance segmentation toolkit based on PaddlePaddle.

paddleseg

!图像分割(表计识别MechanicalIndustryMeter)End-to-end image segmentation kit based on PaddlePaddle.

panoptic-partformer

The First Unified End-to-End System for Panoptic Part Segmentation-ECCV-2022

patchwork-plusplus

!!!可适用于不平坦路面的基于激光点云的快速鲁棒地面分割Fast and robust ground segmentation method using 3D point cloud. @ IROS'22