摘要
通过序列影像获取、卷积神经网络与集成学习的影像场景分类、标志物影像匹配等方法,解决地形平坦而导致GPS或北斗信号受干扰、消弱等环境下的自主导航定位问题,并利用获得的导航定位信息重新规划路径,有效降低人员生命危险,提高任务成功率。
In this paper,we used the methods,such as sequence images acquisition,convolutional neural network and integrated learning image scene classification,marker image matching,to solve the problems of autonomous navigation and positioning under the environment of GPS and BDS being disturbed and weaken.Then,we used navigation and positioning information to replan paths,which could effectively reduce the life-threatening situation and improving the success rate.
出处
《地理空间信息》
2021年第8期101-102,109,I0007,共4页
Geospatial Information
关键词
导航定位
场景分类
影像匹配
navigation positioning
scene classification
image matching
