| 注册
首页|期刊导航|航天器工程|巡视探测任务中复杂地形信息感知与场景建模

巡视探测任务中复杂地形信息感知与场景建模

赵迪 戴志鹏 李世其 纪合超 何宁

航天器工程2019,Vol.5Issue(5):32-38,7.
航天器工程2019,Vol.5Issue(5):32-38,7.DOI:10.3969/j.issn.1673-8748.2019.05.006

巡视探测任务中复杂地形信息感知与场景建模

Perception and Scene Modeling of Complex Terrain Information in Patrol and Exploration Tasks

赵迪 1戴志鹏 2李世其 1纪合超 2何宁2

作者信息

  • 1. 湖北工业大学,武汉 430068
  • 2. 华中科技大学,武汉 430074
  • 折叠

摘要

Abstract

During the lunar probe,many irregular terrain obstacles on the lunar surface (meteorite craters,slopes,lunar rocks,etc.)will affect the mobile performance of the rover,and the lack of direct three-dimensional information about the lunar surface environment affects ground observer’s final decision.In this paper,RGB-D camera is used to obtain the original data,and based on 3D point cloud data,noise filtering and other processing is carried out,combined with the obstacle crossing ability limit of the robot and the improved RANSAC (random sample con-sensus)algorithm,the adaptive reference plane is obtained as the passable region.Finally,DB-SCA (density based spatial clustering of applications with noise)algorithm is used to extract local terrain obstacle information,and rapid 3D scene reconstruction is carried out based on the datum, providing an immediate and rapid 3D environment model of the rover for ground observation, which is verified by simulating the lunar terrain environment experiment.Experimental results show that the algorithm used in this paper can effectively obtain the spatial coordinate information of terrain obstacles,and carry out rapid scene reconstruction,improving greatly the time efficiency.It can provide a reference for the rover to avoid obstacles and provide the ground observer with three-dimensional view.

关键词

点云分割/随机采样一致性算法/密度聚类算法/三维建模

Key words

point cloud segmentation/RANSAC/DBSCAN/3D scene modeling

分类

信息技术与安全科学

引用本文复制引用

赵迪,戴志鹏,李世其,纪合超,何宁..巡视探测任务中复杂地形信息感知与场景建模[J].航天器工程,2019,5(5):32-38,7.

基金项目

国家重大科技专项工程(060601) (060601)

航天器工程

OA北大核心CSCDCSTPCD

1673-8748

访问量0
|
下载量0
段落导航相关论文