Lidar (Light Detection and Ranging) is an active remote-sensing technology that emits pulsed laser light and measures the time-of-flight of returning reflections to compute precise three-dimensional point-cloud representations of the surrounding environment. Operating across wavelengths from near-infrared to ultraviolet, it achieves centimetre-scale spatial accuracy and is robust to many lighting conditions where camera-based systems degrade. Lidar is a foundational sensor modality in autonomous vehicles, aerial surveying, robotics, and spatial-computing applications, commonly fused with IMU, camera, and GNSS data to enable localisation, mapping, and obstacle detection.
Overview
- Lidar functions by projecting narrow laser pulses — typically in the near-infrared spectrum (905 nm or 1550 nm) — toward the scene and timing the round-trip travel of photons with a high-precision clock. The measured time divided by twice the speed of light yields range; combining range with azimuth and elevation angles produces a 3D coordinate for each return pulse.
- Modern systems sample millions of points per second, generating Point Cloud data that capture geometry, surface normals, and in some configurations intensity and multi-echo information.
- Lidar is routinely fused with Inertial Measurement Unit data, GNSS signals, and Camera Sensor imagery in a Sensor Fusion pipeline to compensate for motion distortion and to associate colour or semantic labels with geometric measurements.
- The technology has matured from bulky rotating assemblies to compact solid-state units, driving adoption in consumer vehicles, mobile robots, UAVs, and infrastructure inspection platforms.
Key Components
- Laser Emitter — produces collimated pulses, typically a Infrared Laser diode or fibre laser; wavelength choice affects eye safety ratings and atmospheric penetration.
- Photodetector / Receiver — commonly an Avalanche Photodiode (APD) or Single-Photon Avalanche Diode (SPAD) that converts returning photons to an electrical signal with sub-nanosecond timing resolution.
- Time-of-Flight Sensor / TDC — a Time-to-Digital Converter latches the pulse departure and return timestamps to compute range.
- Scanning Mechanism — mechanical: Rotating Mirror or spinning head; MEMS mirror; optical-phased-array (solid-state); flash lidar (no moving parts, illuminates the full scene simultaneously).
- Signal Processing Pipeline — peak detection, noise filtering, multi-return processing, and range walk correction convert raw waveforms to calibrated point coordinates.
- Point Cloud Processing Stack — downstream software (PCL, Open3D, ROS point-cloud tools) performs registration, segmentation, ground removal, and object detection on the raw 3D data.
Lidar Architectures
- Spinning / Mechanical — 360° horizontal field of view via rotating head; mature, high point density; mechanically complex (e.g., Velodyne HDL-64E).
- MEMS Lidar — micro-mirror steers beam in 2D; smaller, lower power; limited FoV.
- Solid-State / Optical Phased Array — no moving parts; electronically steered beam; rapidly scaling towards automotive cost targets.
- Flash Lidar — full-frame 3D capture in a single pulse; low latency; short range; suited to Augmented Reality and gesture sensing.
- FMCW Lidar — Frequency-Modulated Continuous Wave; simultaneously measures range and radial velocity per point; more robust to interference; growing adoption in automotive.
Applications
- Autonomous Driving — primary long-range obstacle detection and free-space mapping sensor; combined with Radar and Camera Sensor in redundant perception stacks.
- Mobile Robotics / Autonomous Navigation — ground and aerial robots use spinning or solid-state lidar for real-time SLAM and path planning.
- Aerial Surveying (ALS) — airborne lidar mounted on fixed-wing or rotary UAVs captures sub-decimetre Geospatial Data for forestry, archaeology, and infrastructure.
- Terrestrial Laser Scanning (TLS) — static tripod-mounted scanners produce millimetre-accuracy 3D Mapping of buildings, tunnels, and heritage sites.
- Digital Twin Construction — as-built scans feed 3D Scene Reconstruction pipelines feeding BIM and city-scale Spatial Computing platforms.
- Atmospheric Science — Doppler lidar measures wind profiles and aerosol backscatter in the lower troposphere.
- Industrial Inspection — detecting deformation, cracks, and volumetric changes in pipelines, turbine blades, and civil structures.
- Augmented Reality / XR — iPhone and iPad Pro use a flash ToF lidar for rapid room-scale meshing and occlusion in ARKit.
Standards & Context
- IEC 60825 — laser safety classification (Class 1 eye-safe requirements drive 1550 nm wavelength adoption in automotive lidar).
- SAE J3016 — autonomous driving levels context within which lidar sensor requirements are derived.
- ROS 2 / sensor_msgs/PointCloud2 — de facto interface standard for lidar data exchange in robotic middleware.
- LAS / LAZ (ASPRS) — open binary formats for storing aerial and terrestrial lidar point clouds; widely used in Geospatial Data workflows.
- Open3D, PCL (Point Cloud Library) — open-source reference implementations for Point Cloud Processing algorithms.
- ADAS regulation — EU and UNECE regulations for automated driving (UN-R157 ALKS) implicitly require redundant sensing including lidar-class ranging.
- Industry bodies including the LiDAR Association and IEEE Sensors Council drive interoperability and benchmarking standards.