.png){kind=icon}
.png){kind=icon}
Return
Discriminative optimization algorithm with global-local feature for LIDAR point cloud registration
.png){kind=icon}
PRE
AI
DOI：10.1080/01431161.2021.1975843.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The discriminative optimization (DO) algorithm has been successful in three-dimensional LIDAR rigid point cloud registration. However, its feature descriptor is local, which may restrict the robustness of the DO. In this study, for the sake of improving the DO's robustness, we design a global-local feature descriptor, then we weight the global-local descriptor using the prior information from model point cloud. We compare the proposed approach with eight classical point cloud registration methods using the Stanford Bunny, Oxford SensatUrban and Sydney 3D cross source datasets. Experimental results demonstrate that our proposed approach obtains higher registration accuracy and lower registration root-mean-square-error. In addition, our method is competitive in runtime.
Keywords:
TRACKING
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
2.6
Papers:
1.2W
Citations:
2.7W