LIDAR
Awesome LIDAR
¶
A curated list of awesome LIDAR sensors and its applications.
LIDAR is a remote sensing sensor that uses laser light to measure the surroundings in ~cm accuracy. The sensory data is usually referred as point cloud which means set of data points in 3D or 2D. The list contains hardwares, datasets, point cloud-processing algorithms, point cloud frameworks, simulators etc.
Contributions are welcome! Please check out our guidelines.
Conventions¶
- Any list item with an OctoCat :octocat: has a GitHub repo or organization
- Any list item with a RedCircle
has YouTube videos or channel
- Any list item with a Paper
has a scientific paper or detailed description
Manufacturers¶
- Velodyne - Velodyne is a mechanical and solid-state LIDAR manufacturer. The headquarter is in San Jose, California, USA.
- YouTube channel
- ROS driver :octocat:
- Ouster - LIDAR manufacturer, specializing in digital-spinning LiDARs. Ouster is headquartered in San Francisco, USA.
- YouTube channel
- GitHub organization :octocat:
- Livox - LIDAR manufacturer.
- YouTube channel
- GitHub organization :octocat:
- SICK - Sensor and automation manufacturer, the headquarter is located in Waldkirch, Germany.
- YouTube channel
- GitHub organization :octocat:
- Hokuyo - Sensor and automation manufacturer, headquartered in Osaka, Japan.
- YouTube channel
- Pioneer - LIDAR manufacturer, specializing in MEMS mirror-based raster scanning LiDARs (3D-LiDAR). Pioneer is headquartered in Tokyo, Japan.
- YouTube channel
- Luminar - LIDAR manufacturer focusing on compact, auto-grade sensors. Luminar is headquartered Palo Alto, California, USA.
- Vimeo channel
- GitHub organization :octocat:
- Hesai - Hesai Technology is a LIDAR manufacturer, founded in Shanghai, China.
- YouTube channel
- GitHub organization :octocat:
- Robosense - RoboSense (Suteng Innovation Technology Co., Ltd.) is a LIDAR sensor, AI algorithm and IC chipset maufactuirer based in Shenzhen and Beijing (China).
- YouTube channel
- GitHub organization :octocat:
- Ibeo - Ibeo Automotive Systems GmbH is an automotive industry / environmental detection laserscanner / LIDAR manufacturer, based in Hamburg, Germany.
- YouTube channel
- Innoviz - Innoviz technologies / specializes in solid-state LIDARs.
- YouTube channel
- Quanenergy - Quanenergy Systems / solid-state and mechanical LIDAR sensors / offers End-to-End solutions in Mapping, Industrial Automation, Transportation and Security. The headquarter is located in Sunnyvale, California, USA.
- YouTube channel
- Cepton - Cepton (Cepton Technologies, Inc.) / pioneers in frictionless, and mirrorless design, self-developed MMT (micro motion technology) lidar technology. The headquarter is located in San Jose, California, USA.
- YouTube channel
- Blickfeld - Blickfeld is a solid-state LIDAR manufacturer for autonomous mobility and IoT, based in München, Germany.
- YouTube channel
- GitHub organization :octocat:
- Neuvition - Neuvition is a solid-state LIDAR manufacturer based in Wujiang, China.
- YouTube channel
Datasets¶
- Ford Dataset - The dataset is time-stamped and contains raw data from all the sensors, calibration values, pose trajectory, ground truth pose, and 3D maps. The data is Robot Operating System (ROS) compatible.
- Paper
- GitHub repository :octocat:
- Audi A2D2 Dataset - The dataset features 2D semantic segmentation, 3D point clouds, 3D bounding boxes, and vehicle bus data.
- Waymo Open Dataset - The dataset contains independently-generated labels for lidar and camera data, not simply projections.
- Oxford RobotCar - The Oxford RobotCar Dataset contains over 100 repetitions of a consistent route through Oxford, UK, captured over a period of over a year.
- YouTube channel
- Paper
- EU Long-term Dataset - This dataset was collected with our robocar (in human driving mode of course), equipped up to eleven heterogeneous sensors, in the downtown (for long-term data) and a suburb (for roundabout data) of Montbéliard in France. The vehicle speed was limited to 50 km/h following the French traffic rules.
- NuScenes - Public large-scale dataset for autonomous driving.
- Paper
- Lyft - Public dataset collected by a fleet of Ford Fusion vehicles equipped with LIDAR and camera.
- KITTI - Widespread public dataset, pirmarily focusing on computer vision applications, but also contains LIDAR point cloud.
- Semantic KITTI - Dataset for semantic and panoptic scene segmentation.
- YouTube video
- CADC - Canadian Adverse Driving Conditions Dataset - Public large-scale dataset for autonomous driving in adverse weather conditions (snowy weather).
- Paper
- UofTPed50 Dataset - University of Toronto, aUToronto's self-driving car dataset, which contains GPS/IMU, 3D LIDAR, and Monocular camera data. It can be used for 3D pedestrian detection.
- Paper
- PandaSet Open Dataset - Public large-scale dataset for autonomous driving provided by Hesai & Scale. It enables researchers to study challenging urban driving situations using the full sensor suit of a real self-driving-car.
Libraries¶
- Point Cloud Library (PCL) - Popular highly parallel programming library, with numerous industrial and research use-cases.
- GitHub repository :octocat:
- Open3D library - Open3D library contanins 3D data processing and visualization algorithms. It is open-source and supports both C++ and Python.
- GitHub repository :octocat:
- YouTube channel
- PyTorch Geometric
- A geometric deep learning extension library for PyTorch.
- GitHub repository :octocat:
- PyTorch3d - PyTorch3d is a library for deep learning with 3D data written and maintained by the Facebook AI Research Computer Vision Team.
- GitHub repository :octocat:
- Kaolin - Kaolin is a PyTorch Library for Accelerating 3D Deep Learning Research written by NVIDIA Technologies for game and application developers.
- GitHub repository :octocat:
- Paper
- PyVista - 3D plotting and mesh analysis through a streamlined interface for the Visualization Toolkit.
- GitHub repository :octocat:
- Paper
- pyntcloud - Pyntcloud is a Python 3 library for working with 3D point clouds leveraging the power of the Python scientific stack.
- GitHub repository :octocat:
Frameworks¶
- Autoware - Popular framework in academic and research applications of autonomous vehicles.
- GitLab repository :octocat:
- Paper
- Baidu Apollo - Apollo is a popular framework which accelerates the development, testing, and deployment of Autonomous Vehicles.
- GitHub repository :octocat:
- YouTube channel
Algorithms¶
Basic matching algorithms¶
- Iterative closest point
- The must-have algorithm for feature matching applications (ICP).
- Normal distributions transform
- More recent massively-parallel approach to feature matching (NDT).
Semantic segmentation¶
- RangeNet++
- Fast and Accurate LiDAR Sematnic Segmentation with fully convolutional network.
- GitHub repository :octocat:
- YouTube video
- PolarNet
- An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation.
- GitHub repository :octocat:
- YouTube video
- Frustum PointNets
- Frustum PointNets for 3D Object Detection from RGB-D Data.
- GitHub repository :octocat:
- Study of LIDAR Semantic Segmentation - Scan-based Semantic Segmentation of LiDAR Point Clouds: An Experimental Study IV 2020.
- Paper
- GitHub repository :octocat:
Simultaneous localization and mapping SLAM and LIDAR-based odometry and or mapping LOAM¶
- LOAM J. Zhang and S. Singh
- LOAM: Lidar Odometry and Mapping in Real-time.
- LeGO-LOAM :octocat: - A lightweight and ground optimized lidar odometry and mapping (LeGO-LOAM) system for ROS compatible UGVs.
- YouTube video
- Cartographer :octocat: - Cartographer is ROS compatible system that provides real-time simultaneous localization and mapping (SLAM) in 2D and 3D across multiple platforms and sensor configurations.
- YouTube video
- SuMa++
- LiDAR-based Semantic SLAM.
- GitHub repository :octocat:
- YouTube video
- OverlapNet
- Loop Closing for LiDAR-based SLAM.
- GitHub repository :octocat:
- YouTube video
- LIO-SAM
- Tightly-coupled Lidar Inertial Odometry via Smoothing and Mapping.
- GitHub repository :octocat:
- YouTube video
Object detection and object tracking¶
- Learning to Optimally Segment Point Clouds
- By Peiyun Hu, David Held, and Deva Ramanan at Carnegie Mellon University. IEEE Robotics and Automation Letters, 2020.
- YouTube video
- GitHub repository :octocat:
- Leveraging Heteroscedastic Aleatoric Uncertainties for Robust Real-Time LiDAR 3D Object Detection
- By Di Feng, Lars Rosenbaum, Fabian Timm, Klaus Dietmayer. 30th IEEE Intelligent Vehicles Symposium, 2019.
- YouTube video
- What You See is What You Get: Exploiting Visibility for 3D Object Detection
- By Peiyun Hu, Jason Ziglar, David Held, Deva Ramanan, 2019.
- YouTube video
- GitHub repository :octocat:
Simulators¶
- CoppeliaSim - Cross-platform general-purpose robotic simulator (formerly known as V-REP).
- YouTube channel
- OSRF Gazebo - OGRE-based general-purpose robotic simulator, ROS/ROS2 compatible.
- GitHub repository :octocat:
- CARLA - Unreal Engine based simulator for automotive applications. Compatible with Autoware, Baidu Apollo and ROS/ROS2.
- GitHub repository :octocat:
- YouTube channel
- LGSVL - Unity Engine based simulator for automotive applications. Compatible with Autoware, Baidu Apollo and ROS/ROS2.
- GitHub repository :octocat:
- YouTube channel
Others¶
- Pointcloudprinter :octocat: - A tool to turn point cloud data from aerial lidar scans into solid meshes for 3D printing.
- Pcx :octocat: - Point cloud importer/renderer for Unity.
- Bpy :octocat: - Point cloud importer/renderer/editor for Blender, Point Cloud visualizer.
- Semantic Segmentation Editor :octocat: - Point cloud and image semantic segmentation editor by Hitachi Automotive And Industry Laboratory.
- Photogrammetry importer :octocat: - Blender addon to import reconstruction results of several libraries.