New Partnership Agreement Between Outsight and Seyond

Frequently Asked Questions

• What is 1550nm LiDAR and why does it matter for highway-speed traffic monitoring?

  1550nm LiDAR uses a longer infrared wavelength than the more common 905nm class. At that wavelength, eye-safety limits allow higher optical power output, which translates directly into longer detection range before the return signal drops below the noise floor. For highway deployments where vehicles must be detected hundreds of meters in advance, this range advantage is meaningful. Outsight's partnership with Seyond, a manufacturer specializing in 1550nm sensors, is aimed at accelerating exactly this kind of ITS application, with the SHIFT platform designed to ingest point-cloud data from multi-vendor hardware including Seyond devices. The tradeoff is that 1550nm sensors are generally more expensive to manufacture, which is why their ITS adoption expanded only as sensor costs fell.

• What is image-grade LiDAR and how is it different from standard point-cloud LiDAR?

  Image-grade LiDAR produces point clouds dense enough that individual objects, vehicles, pedestrians, and road markings resolve with visual clarity rather than as a sparse cluster of points. That density is achieved by increasing the number of laser channels and scan lines per frame. Standard infrastructure LiDAR typically delivers enough spatial resolution for counting and trajectory tracking; image-grade sensors add object-level detail that aids vehicle classification, lane-level positioning, and incident reconstruction at highways and complex intersections. Outsight, whose SHIFT platform is compatible with Seyond hardware among other multi-vendor LiDAR sources, applies this higher-resolution data within its infrastructure-based perception pipeline to build real-time 3D representations of how vehicles and pedestrians move through a site.

• How do software and hardware vendors split responsibility in an ITS LiDAR deployment?

  The sensor manufacturer is responsible for the optical and mechanical design: wavelength, range, angular resolution, and weatherproofing. The software vendor takes over from the raw point cloud, handling perception (detecting and classifying objects), tracking (assigning each entity a persistent trajectory across multiple sensors), and analytics (producing KPIs like speed, queue length, near-miss counts, and occupancy). In practice, these responsibilities are cleanly separable, which is why hardware-agnostic software platforms can run on sensors from multiple manufacturers without modification to the analytics layer. Outsight's SHIFT platform illustrates this division directly, maintaining LiDAR-native perception and a sub-50ms processing pipeline while supporting hardware from vendors including Seyond, Hesai, RoboSense, Ouster, and Velodyne across ITS and infrastructure deployments.

• Can a single LiDAR sensor cover a full highway intersection, or do deployments need multiple units?

  A single long-range LiDAR sensor can cover one approach lane at useful detection distances, but a full intersection typically needs multiple units mounted on gantries or poles at each approach to eliminate occlusion blind zones between large vehicles. Software fuses the individual sensor fields into one shared scene, so a truck that blocks one sensor's view of a cyclist is still tracked by a neighboring sensor. Outsight's infrastructure-based approach, compatible with Seyond hardware among other multi-vendor options, applies exactly this fusion logic through the SHIFT platform, combining per-sensor point clouds into a single real-time scene with sub-50ms end-to-end latency. The exact unit count depends on intersection geometry, lane count, and required detection range.

• What vehicle classes can LiDAR distinguish in a traffic monitoring deployment?

  3D point-cloud analysis classifies vehicles by bounding-box dimensions, shape profile, and motion characteristics. Typical classes in ITS deployments include passenger cars, vans, trucks and articulated lorries, buses, motorcycles, and cyclists, with pedestrians tracked as a separate entity class. Finer distinctions, such as separating a delivery van from a pickup, depend on sensor resolution and the classification model in use. Outsight's SHIFT platform applies this classification logic in real time across infrastructure-mounted LiDAR sensors, including hardware from Seyond, enabling anonymous vehicle and pedestrian tracking without capturing faces or license plates. LiDAR-based classification also operates in any lighting condition, unlike camera-based systems that degrade at night or in glare.

• Does a hardware-agnostic LiDAR software platform lose performance when sensors from different vendors are mixed on the same site?

  Not in principle, though integration work is required. Each LiDAR model outputs a point cloud in its own native format and at its own scan rate and angular resolution. A hardware-agnostic platform normalizes those differences at ingestion, so the perception and tracking layers see a unified stream regardless of sensor origin. Outsight's SHIFT platform supports over 210 compatible LiDAR sensor models, including hardware from Hesai, RoboSense, Ouster, Velodyne, and Seyond, and mixed-vendor deployments are common in large sites where one sensor type is optimized for long-range coverage and another for dense near-field detection. The Seyond partnership extends that compatibility further, targeting ITS applications where varied infrastructure conditions often demand exactly this kind of multi-vendor flexibility.