摘要
针对石油平台导管架水下清洗机器人姿态定位问题,提出一种新型的视觉惯导组合导航系统。该系统利用Canny算子与Hough变换识别圆管特征直线,并通过建立机器人空间模型,解算机器人空间位置姿态角。针对视觉惯导系统模型,设计基于四元数的扩展卡尔曼滤波,实现视觉与惯导信息融合,减小空间角度的累计误差,提高组合导航精度。研究结果表明:本文提出的机器人组合导航系统角度精度高,其精确度及实时性满足了机器人实际应用需求。
For the localization of the underwater cleaning robot’s localization,an innovative vision-inertial navigation system was put forward.By recognizing pipe line in Canny/Hough and establishing robot’s spatial model,the attitude angles were solved.For vision-inertial system model,the extended Kalman filter bases on quaternion was designed to fuse vision/inertial data and reduce accumulative errors,which improves the system accuracy.The results show that this attitude accuracy is high,which indicates that the real-time performance and the precision of the navigation system can meet the practical demand.
作者
江平
杨灿军
寿志成
陈燕虎
范锦昌
黄政明
魏谦笑
JIANG Ping;YANG Canjun;SHOU Zhicheng;CHEN Yanhu;FAN Jinchang;HUANG Zhengming;WEI Qianxiao(State Key Laboratory of Fluid Power and Mechatronic Systems, Zhejiang University, Hangzhou 310027, China;Shanghai Branch of China National Offshore Oil Corporation, Shanghai 200335, China)
出处
《中南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2018年第12期2946-2952,共7页
Journal of Central South University:Science and Technology
基金
浙江省公益技术应用研究计划项目(2016C33057)
中国海洋石油总公司科技项目(CNOOC-KJ125 ZDXM 13 LTD NFZB 2015-04)~~
关键词
姿态定位
视觉导航
组合导航
图像识别
扩展卡尔曼滤波
attitude and location
vision navigation system
integrated navigation
image recognition
extended Kalman filter
