Abstract
Registration of laser scans, or point clouds in general, is a crucial step of localization and mapping with mobile robots or in object modeling pipelines. A coarse alignment of the point clouds is generally needed before applying local methods such as the Iterative Closest Point (ICP) algorithm. We propose a feature-based approach to point cloud registration and evaluate the proposed method and its individual components on challenging real-world datasets. For a moderate overlap between the laser scans, the method provides a superior registration accuracy compared to state-of-the-art methods including Generalized ICP, 3D Normal-Distribution Transform, Fast Point-Feature Histograms, and 4-Points Congruent Sets. Compared to the surface normals, the points as the underlying features yield higher performance in both keypoint detection and establishing local reference frames. Moreover, sign disambiguation of the basis vectors proves to be an important aspect in creating repeatable local reference frames. A novel method for sign disambiguation is proposed which yields highly repeatable reference frames.
Highlights
Point cloud registration has many applications including mobile robotics, object modeling, and object recognition and pose estimation
Prior to evaluating the registration method as a whole following the protocol of [4], we evaluate keypoint detection and local reference frames using a small number of laser scans and fix their parameters
We have not found any to be significantly better than the others across all scales, both in keypoint detection and local reference frames, and only the unit weights are considered further
Summary
Point cloud registration has many applications including mobile robotics, object modeling, and object recognition and pose estimation. It is a crucial step of the most commonly used methods for Simultaneous Localization and Mapping (SLAM), whether operating on the data from laser scanners or consumer-electronics RGB-D sensors, which have become widely available. A variant of the Iterative Closest Points (ICP) algorithm is often employed to solve the task— see [1, 2] for the seminal papers on its point-to-point and point-to-plane formulations, respectively, or [3] for a generalization of these two methods. Despite many advantages of the algorithm, including real-time operation in some settings, the ICP algorithm has several drawbacks. As shown by [4], an inaccurate
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have