Paper Info
Title
Performance Evaluation of Querying Point Clouds in RDBMS.
Abstract
A point cloud is a data set including a number of 3-dimensional point data. Mobile robots such as autonomous vehicles require point clouds for localization, which is a technique for estimating self-position. Localization is a kind of similarity searches between the current point cloud from on-board sensors and historical point clouds observed in advance. Thus, how to manage historical point clouds is quite important. But, some experimental systems for mobile robots do not care about data management of historical point clouds. They are usually stored in text files and are fully loaded on memory at the time of startup. We believe that historical point clouds should be managed by DBMS and loaded by queries if need arises. Before challenging this research theme, in this paper, we investigate performance of querying point clouds stored in RDBMS (PostgreSQL and PostGIS). We measure execution time for range query and rectangular query, and performance of B-tree index and R-tree index.
Year
DOI
Venue
2019
10.1109/BIGCOMP.2019.8679485
BigComp
Keywords
Field
DocType
Three-dimensional displays,Indexes,Laser radar,Mobile robots,Time measurement,Autonomous vehicles,Roads
Computer science,Range query (data structures),Lidar,Execution time,Relational database management system,Point cloud,Data management,Mobile robot,Database
Conference
ISSN
ISBN
Citations 
2375-933X
978-1-5386-7789-6
0
PageRank 
References 
Authors
0.34
0
4
Name
Order
Citations
PageRank
Gen Ikawa100.34
Yousuke Watanabe282.06
Shunya Yamada300.34
Hiroaki Takada460887.55