摘要
设计了一种基于多传感器融合的室内建图方法,能够实现在包含玻璃物体的室内环境进行地图构建。主要利用激光雷达采集的数据和双目相机的数据进行特征级融合,再通过扩展卡尔曼滤波进行位姿融合;依据激光雷达对数据进行采集,利用参考阈值对疑似点进行判定;进一步使用双目视觉进行识别判定,实现多传感器融合建图。该方法可以提高局部定位能力和全局定位能力,解决传统Gmapping算法中对玻璃无法准确识别的问题。
In this paper, a multi-sensor fusion-based indoor mapping method is designed that enables map construction in indoor environments containing glass objects. The data collected by LIDAR and binocular camera are mainly used for feature-level fusion, and then the bit-pose fusion is performed by extended Kalman filtering. The suspected points are determined by using reference thresholds, and are further recognized and determined by using binocular vision to build the map with multi-sensor fusion. This method can improve the local localization ability and global localization ability, and solve the problem that the traditional Gmapping algorithm cannot accurately identify the glass.
作者
郭辰
万里
雍雨晨
王伟
胡佳飞
GUO Chen;WAN Li;YONG Yuchen;WANG Wei;HU Jiafei(Undergraduated School,National University of Defense Technology,Changsha 410003,China;Engineering College of Northeast Agricultural University,Harbin150080,China;College of Intelligence Science and Technology,National University of Defense Technology,Changsha 410003,China)
出处
《仪表技术》
2022年第6期39-43,共5页
Instrumentation Technology
