Fast and robust 3D feature extraction from sparse point clouds

Matching 3D point clouds, a critical operation in map building and localization, is difficult with Velodyne-type sensors due to the sparse and non-uniform point clouds that they produce. Standard methods from dense 3D point clouds are generally not effective. In this paper, we describe a feature-based approach using Principal Components Analysis (PCA) of neighborhoods of points, which results in mathematically principled line and plane features. The key contribution in this work is to show how this type of feature extraction can be done efficiently and robustly even on non-uniformly sampled point clouds. The resulting detector runs in real-time and can be easily tuned to have a low false positive rate, simplifying data association. We evaluate the performance of our algorithm on an autonomous car at the MCity Test Facility using a Velodyne HDL-32E, and we compare our results against the state-of-the-art NARF keypoint detector.

Authors

• Jacopo Serafin
• Edwin Olson
• Giorgio Grisetti

Keywords

• Three-dimensional displays
• Feature extraction
• Robustness
• Detectors
• Laser radar
• Simultaneous localization and mapping
• Point Cloud
• Robust Features
• 3D Features
• Sparse Point Cloud
• Robust Feature Extraction
• 3D Feature Extraction
• Self-driving
• Neighboring Points
• 3D Point Cloud
• Line Features
• Dense Point Cloud
• Planar Features
• Good Characteristics
• Autonomous Vehicles
• RGB Images
• Standard Light
• Mobile Robot
• Feature Calculation
• External Noise
• RGB Camera
• 3D Plane
• Scale-invariant Feature Transform
• 3D Line
• Surface Normals
• Integral Image
• Spherical Image
• Input Point Cloud
• Speeded Up Robust Features
• Histogram Features