摘要
为了保证电力系统安全、稳定地运行,该研究设计了一种基于自适应语义地图的电力巡检路径规划算法。首先,通过ORB-SLAM2视觉框架获取电力巡检图像,在提取图像的ORB特征后,利用YOLACT网络分割图像,得到实例掩码;其次,利用深度信息与自适应区域增长算法,完成语义标注;再次,基于栅格及八叉树结构构建电力巡检环境自适应语义地图;最后,以上述构建的语义地图为环境基础,将分流蚁模拟退火算法与混沌扰动初始化算法相结合,搜寻出一条最佳的巡检路径。实验结果表明:该算法构建的自适应语义地图可准确描述电力巡检环境,且电力巡检路径短,路径规划效率高。
In order to ensure the safe and stable operation of the power system,this study designs a power inspection path planning algorithm based on adaptive semantic map.Firstly,the power inspection image is obtained using the ORB-SLAM2 visual framework.After extracting the ORB features of the image,the YOLACT network is used to segment the image and obtain the instance mask.Secondly,deep information and adaptive region growth algorithm are used to complete semantic annotation.Thirdly,an adaptive semantic map of power inspection environment is constructed based on grid and Octree structure.Finally,based on the semantic map constructed above,the diffluent ant simulated annealing algorithm is combined with the chaos perturbation initialization algorithm to find an optimal patrol path.The experiment results show that the adaptive semantic map constructed by this algorithm can accurately describe the power inspection environment,and the power inspection path is short and the path planning efficiency is high.
作者
李琰
LI Yan(Guangxi Vocational University of Agriculture,Nanning 530028,China)
出处
《信息技术》
2024年第10期168-174,共7页
Information Technology
关键词
语义地图
自适应
电力巡检
路径规划
改进蚁群算法
semantic map
adaptive
power inspection
path planning
improved ant colony algorithm
