基于ORB-SLAM3的多旋翼室内定位方法研究

Research on indoor positioning method of multiple rotors method based on ORB-SLAM 3

多旋翼无人机作为最常见的空中机器人,在电力巡检、气象探测、公路巡视、灾后救援、航拍等诸多领域发挥着重要的作用。在室内GNSS信号拒止环境中,基于视觉或激光雷达前端的同时定位与建图(SLAM)方法的实时定位方法越来越受到关注。文章利用数据集验证了紧耦合的ORB-SLAM 3双目/IMU算法在不同场景、不同光照以及不同运动状态下无人机的定位精度以及鲁棒性。实验表明,算法可实现厘米级的全局定位精度,并且在快速运动导致的图像模糊的情况下,仍能保持精确地跟踪,具有较好的鲁棒性。

Multi-rotor UAVs,as the most common aerial robots,play an important role in many fields such as power inspection,weather detection,highway patrol,post-disaster rescue,and aerial photography.However,in indoor environments such as GNSS signal rejection,real-time localization methods based on vision or LIDAR front-end simultaneous localization and map building(SLAM)methods have received increasing attention.In this paper,we utilize a dataset to validate the positioning accuracy as well as the robustness of the tightly coupled ORB-SLAM 3 binocular/IMU algorithm for UAVs in different motion states.Experiments show that the algorithm can achieve centimeter-level global localization accuracy and maintain accurate tracking with good robustness under image blurring caused by fast motion.