Book Chapter10.1007/978-3-319-29246-5_4
Point Cloud Registration
Martin Weinmann
- 01 Jan 2016
- pp 55-110
8
TL;DR: The results clearly reveal that the further consideration of both a correspondence weighting based on point quality measures and a selection of an appropriate feature detector–descriptor combination may result in significant advantages with respect to robustness, efficiency, and accuracy.
read more
Abstract: In this chapter, we focus on keypoint-based point cloud registration which has proven to be among the most efficient strategies for aligning pairs of overlapping scans. We present a novel and fully automated framework which consists of six components addressing (i) the generation of 2D image representations in the form of range and intensity images, (ii) point quality assessment, (iii) feature extraction and matching, (iv) the forward projection of 2D keypoints to 3D space, (v) correspondence weighting, and (vi) point cloud registration. For the respective components, we take into account different approaches and our main contributions address the issue of how to increase the robustness, efficiency, and accuracy of point cloud registration by either introducing further constraints (e.g., addressing a correspondence weighting based on point quality measures) or replacing commonly applied approaches by more promising alternatives. The latter may not only address the involved strategy for point cloud registration, but also the involved approaches for feature extraction and matching. In a detailed evaluation, we demonstrate that, instead of directly aligning sets of corresponding 3D points, a transfer of the task of point cloud registration to the task of solving the Perspective-n-Point (PnP) problem or to the task of finding the relative orientation between sets of bearing vectors offers great potential for future research. Furthermore, our results clearly reveal that the further consideration of both a correspondence weighting based on point quality measures and a selection of an appropriate feature detector–descriptor combination may result in significant advantages with respect to robustness, efficiency, and accuracy.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Registration of Laser Scanning Point Clouds: A Review
TL;DR: A comprehensive review of LiDAR data registration in the fields of photogrammetry and remote sensing is presented, and the lack of standard data and unified evaluation systems is identified as a factor limiting objective comparison of different methods.
251
Accuracy and reliability evaluation of 3D-LS for the discontinuity orientation identification with different registration/georeferencing modes
TL;DR: This study examined the accuracy and reliability of the scanner’s built-in direction system, and observed that orientations captured by 3D-LS can be more accurate with RTK registration/georeferencing than manual survey.
5
Rethinking Point Cloud Registration as Masking and Reconstruction
Guangyan Chen,Meiling Wang,Li Yuan,Yi Yang,Yufeng Yue +4 more
- 01 Oct 2023
TL;DR: A generic and concise auxiliary training network, the Masked Reconstruction Auxiliary Network (MRA), is proposed, which reconstructs the complete point cloud by separately using the encoded features of each point cloud obtained from the backbone to capture fine-grained geometric details and the overall structures of point cloud pairs.
4
A review of rigid point cloud registration based on deep learning
Lei Chen,Changzhou Feng,Yun Fei Ma,Yikai Zhao,Chaorong Wang +4 more
TL;DR: This review summarizes the point cloud registration technology based on deep learning, and outlooks the current challenges and future research directions of deep learning-based point cloud registration.
4
Effect of different registration methods on precision of orientation based on RTK registration/georeferencing mode
TL;DR: It is found that the common and the optimization registration method can meet the project’s engineering requirements, and the optimized registration method improved accuracy in the dip direction by approximately 1°.
3
References
Distinctive Image Features from Scale-Invariant Keypoints
TL;DR: This paper presents a method for extracting distinctive invariant features from images that can be used to perform reliable matching between different views of an object or scene and can robustly identify objects among clutter and occlusion while achieving near real-time performance.
Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography
TL;DR: New results are derived on the minimum number of landmarks needed to obtain a solution, and algorithms are presented for computing these minimum-landmark solutions in closed form that provide the basis for an automatic system that can solve the Location Determination Problem under difficult viewing.
A method for registration of 3-D shapes
Paul J. Besl,H.D. McKay +1 more
TL;DR: In this paper, the authors describe a general-purpose representation-independent method for the accurate and computationally efficient registration of 3D shapes including free-form curves and surfaces, based on the iterative closest point (ICP) algorithm, which requires only a procedure to find the closest point on a geometric entity to a given point.
20.6K
•Book
Multiple view geometry in computer vision
Richard Hartley,Andrew Zisserman +1 more
- 01 Jan 2000
TL;DR: In this article, the authors provide comprehensive background material and explain how to apply the methods and implement the algorithms directly in a unified framework, including geometric principles and how to represent objects algebraically so they can be computed and applied.
20.1K
Object recognition from local scale-invariant features
David G. Lowe
- 20 Sep 1999
TL;DR: Experimental results show that robust object recognition can be achieved in cluttered partially occluded images with a computation time of under 2 seconds.