Litcius/Paper detail

Centralized RANSAC-Based Point Cloud Registration With Fast Convergence and High Accuracy

Kuo‐Liang Chung, Wei-Tai Chang

2024IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing15 citationsDOIOpen Access PDF

Abstract

For point cloud registration, the purpose of this paper is to propose a novel centralized RANSAC (C-RANSAC) registration with fast convergence and high accuracy. In our algorithm, the novel contributions are (1) the proposal of a scale histogram-based outlier removal to delete outliers from the initial line vector set <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathbf {L}$</tex-math></inline-formula> for constructing a reduced line vector set <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathbf {L}_{red}$</tex-math></inline-formula> ; (2) the handshake cooperation between the host RANSAC (H-RANSAC) only working on L and the local RANSAC (LCL-RANSAC) only working on <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\mathbf {L}_{red}$</tex-math></inline-formula> ; (3) in each handshake process, after receiving the global registration solution and the global iteration number <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$x_{H}$</tex-math></inline-formula> from H-RANSAC, LCL-RANSAC uses the received global solution as the initial solution of the modified TEASER++ (M-TEASER++) method to calculate its first local registration solution. If the first local registration solution satisfies the global iteration number inheritance condition, LCL-RANSAC directly sends the accumulated iteration number, <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$x_{H} + 1$</tex-math></inline-formula> , and the first local solution back to H-RANSAC; otherwise, LCL-RANSAC iteratively refines its local solution using the M-TEASER++ method, and then sends the resultant local solution and the required local iteration number <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$x_{LCL}$</tex-math></inline-formula> to H-RANSAC for updating the global solution, the global iteration number to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$x_{H} := x_{H} + x_{LCL}$</tex-math></inline-formula> , and the global confidence level. Due to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$|L_{red}| \ll |L|$</tex-math></inline-formula> and employing the global iteration number inheritance condition test into our algorithm, we have conducted extensive experiments on testing point cloud pairs to show the registration accuracy and execution time merits of our algorithm relative to the state-of-the-art methods.

Topics & Concepts

RANSACConvergence (economics)MathematicsAlgorithmOutlierNotationPoint cloudArtificial intelligenceComputer scienceImage (mathematics)ArithmeticEconomic growthEconomics3D Surveying and Cultural HeritageRobotics and Sensor-Based LocalizationRemote Sensing and LiDAR Applications
Centralized RANSAC-Based Point Cloud Registration With Fast Convergence and High Accuracy | Litcius