摘要
SLAM是解决在未知环境进行定位建图的有效工具,能够实现自主定位和导航。针对同时定位与建图系统(ORB-SLAM3)在军用无人机应用路径规划时出现的轨迹丢失问题,提出一种具备重新初始化、关键帧优化、回环检测、轨迹拼接以及地图恢复能力的视觉SLAM系统。该方法通过在轨迹丢失前后进行地图点的保存和匹配作为轨迹拼接的依据,并借助图优化的方式进行丢失前后两段轨迹的拼接工作,最后在拼接后的轨迹中继续执行SLAM任务,完成无人机的路径规划任务。实验结果表明在使用ZED2相机时,利用所提出的SLAM技术可以完成在高空时的定位和建图工作,同时在和GPS的比较中也得到了平均误差为0.8%,能够保证当系统应用于无人机上时在无GPS的情况下能够定位自身的位置,从而完成轨迹规划工作。
SLAM is an effective tool for location mapping in unknown environments,and can achieve autonomous positioning and navigation.Aiming at the problem of trajectory loss in the application of simultaneous localization and mapping system(ORB-SLAM3)in military UAV path planning,a visual SLAM system with re-initialization,key frame optimization,loop detection,trajectory stitching and map recovery capabilities is proposed.The method uses the preservation and matching of map points before and after the trajectory loss as the basis for trajectory stitching,and uses graph optimization to stitch the two trajectories before and after the loss.Finally,the SLAM task is continued in the stitched trajectory to complete the path planning task of the UAV.The experimental results show that when using the ZED2 camera,the SLAM technology proposed can complete the positioning and mapping work at high altitude.At the same time,the average error is 0.8%in comparison with GPS,which can ensure that when the system is applied to the UAV,it can locate its own position without GPS,so as to complete the trajectory planning work.
作者
史佳锋
詹恒泽
孙亚强
历宁
周怡
张文清
Shi Jiafeng;Zhan Hengze;Sun Yaqiang;Li Ning;Zhou Yi;Zhang Wenqing(Guangdong Institute of Artificial Intelligence and Advanced Computing,Guangzhou 510700,China;Institute of Automation,Chinese Academy of Sciences,Beijing 100049,China)
出处
《机电工程技术》
2023年第9期152-156,共5页
Mechanical & Electrical Engineering Technology
基金
广州市科技计划项目(202201000009)。
