Line laser point cloud segmentation based on the combination of RANSAC and region growing

RANdom SAmpling Consensus (RANSAC) and region growing algorithms are widely used in image processing and point cloud segmentation, but the RANSAC algorithm used for point cloud segmentation will cause insufficient segmentation. The region growing algorithm can divide point cloud data into points bas...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Chinese Control Conference s. 6324 - 6328
Hlavní autoři: Yuan, Henan, Sun, Wei, Xiang, Tianyuan
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: Technical Committee on Control Theory, Chinese Association of Automation 01.07.2020
Témata:
Clustering algorithms
Data models
Feature extraction
Image segmentation
line laser
Mathematical model
point cloud segmentation
RANSAC
region growing
Robustness
Three-dimensional displays
ISSN:1934-1768
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:RANdom SAmpling Consensus (RANSAC) and region growing algorithms are widely used in image processing and point cloud segmentation, but the RANSAC algorithm used for point cloud segmentation will cause insufficient segmentation. The region growing algorithm can divide point cloud data into points based on the curvature and normal characteristics of the point cloud. Multiple clusters are easy to be over-segmented. To solve this problem, this paper proposes to use the RANSAC algorithm to perform coarse segmentation to segment the point cloud data into a foreground point cloud with more geometric features and a background point cloud that is only a plane. Then use the region growing algorithm. The foreground point cloud is finely segmented. Besides, the curvature characteristics of the region growing process are used to optimize the plane extraction of the RANSAC algorithm. The experimental results show that this method can reduce over-segmentation to a certain extent and significantly improve the speed of the algorithm.
ISSN:1934-1768
DOI:10.23919/CCC50068.2020.9188506