微型无人机视觉定位与环境建模研究  

Visual Localization and Environment Mapping for Micro Aerial Vehicles

作　　者：吕科[1] 施泽南 李一鹏[2] 

机构地区：[1]中国科学院大学工程科学学院 [2]清华大学信息科学技术学院

出　　处：《电子科技大学学报》2017年第3期543-548,共6页Journal of University of Electronic Science and Technology of China

基　　金：国家自然科学基金面上项目(61271268)

摘　　要：同步定位与环境建模(SLAM)是实现无人机自主飞行和智能导航的关键技术。该文提出了适用于微型无人机的视觉定位与环境建模方法,针对RGB-D传感器使用point-plane ICP点云匹配算法实现视觉自主定位,利用并行计算加速以满足无人机控制的实时性要求;采用TSDF算法融合多帧观测的点云数据,实现了无人机对未知目标环境区域的模型重建;将视觉SLAM系统与无人机载IMU传感器融合,进一步提升了自主定位和建模精度。实际搭建了微型无人机视觉与环境建模验证系统,室内环境下可以达到0.092 m的定位偏差和60 Hz的更新速率,满足了无人机控制的精度和实时性要求,验证了该方法的有效性。Simultaneous localization and mapping （SLAM） is a key technology for unmanned aerial vehicle （UAV） autonomous flight and navigation. This paper presents a visual localization and environment mapping method for micro UAV. In the method, the point-plane ICP point cloud matching algorithm is used to achieve visual localization based on RGB-D sensor, the parallel computing acceleration is applied to meet the real-time requirements of UAV control, then the truncated signed distance function （TSDF） algorithm is used for multi-frame point cloud data fusion to achieve environment model reconstruction for an unknown environment, and at last the fusion of visual SLAM system and UAV IMU sensor data is applied to enhance the precision of autonomous localization and model reconstruction. Finally, we set up a micro UAV system with full visual SLAM algorithm and test the whole system in a typical indoor environment. The result shows a 0.092 m positioning deviation and 60 Hz update rate, capable for real-time UAV controlling accuracy requirement, thus verifying the effectiveness of the method.

关 键 词：环境建模 微型无人机 多传感器 同步定位与环境建模 视觉定位 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]