Data Registration with Ground Points for Roadside LiDAR Sensors

• Main Authors: Rui Yue, Hao Xu, Jianqing Wu, Renjuan Sun, Changwei Yuan
• Format: Article
• Language: English
• Published: MDPI AG 2019-06-01
• Series: Remote Sensing
• Subjects: data registration; Smart Traffic Infrastructure; ground points; optimization
• Online Access: https://www.mdpi.com/2072-4292/11/11/1354

The Light Detection and Ranging (LiDAR) sensors are being considered as new traffic infrastructure sensors to detect road users’ trajectories for connected/autonomous vehicles and other traffic engineering applications. A LiDAR-enhanced traffic infrastructure system requires multiple LiDAR sensors around intersections, along with road segments, which can provide a seamless detection range at intersections or along arterials. Each LiDAR sensor generates cloud points of surrounding objects in a local coordinate system with the sensor at the origin, so it is necessary to integrate multiple roadside LiDAR sensors’ data into the same coordinate system. None of existing methods can integrate the data from roadside LiDAR sensors, because the extensive detection range of roadside sensors generates low-density cloud points and the alignment of roadside sensors is different from mapping scans or autonomous sensing systems. This paper presents a method to register datasets from multiple roadside LiDAR sensors. This approach innovatively integrates LiDAR datasets with 3D cloud points of road surface and 2D reference point features, so the method is abbreviated as RGP (Registration with Ground and Points). The RGP method applies optimization algorithms to identify the optimized linear coordinate transformation. This research considered the genetic algorithm (global optimization) and the hill climbing algorithm (local optimization). The performance of the RGP method and the different optimization algorithms was evaluated with field LiDAR sensors data. When the developed process can integrate data from roadside sensors, it can also register LiDAR sensors’ data on an autonomous vehicle or a robot.