| Replacing manual inspection of high voltage overhead transmission lines with a power inspection robot has become the primary solution to ensure circuit maintenance is both safe,efficient and intelligent.At present,the main functions realized by electric inspection robots include climbing over obstacles,walking on the line,image transmission of overhead transmission lines,abnormal diagnosis of heating or broken strand,etc.The existing electric inspection robots have not been able to realize the inspection function based on three-dimensional environment reconstruction of overhead transmission lines,that is,to create three-dimensional virtual space of working environment in real time.At the same time,robot ontology and transmission line model are integrated to provide assistance and constraints for robot inspection,maintenance and teleoperation,which is very important for the online maintenance of transmission line patrol robot.To ensure the robot’s replacement of maintenance personnel is successful,a system must be established that can provide large-scale three-dimensional real-time outdoor reconstruction and ensure the accuracy of the reconstruction is consistent with the operating conditions.With the development and marketization of Li DAR and relatively low price RGB-D depth cameras in recent years,RGB-D depth cameras have become a device of great concern for real-time 3D reconstruction,gradually developing applications in robot control systems.This paper proposes an overhead transmission line operation robot control system to verify the feasibility of robot operation control simulation based on 3D reconstruction system and ROS operating system.The main contents of the robot control system for overhead transmission line operation are: To address some issues with intelligence and spatial adaptability of current power inspection robots,this paper proposes an alternative.(1)Transmission line 3D reconstruction module.This paper adopts a 3D reconstruction method based on depth cameras to address the digitalization of overhead transmission line operating environments,comprising five steps: point cloud acquisition,preprocessing,registration,fusion,and surface reconstruction.RGB-D depth camera transforms the operating environment of the overhead transmission line operating robot into depth information,which is then transformed into threedimensional point cloud through the point cloud reconstruction method.Through the point cloud preprocessing and matching,the required three-dimensional point cloud format of the inspection robot operating scene is obtained.At long last,the TSDF surface reconstruction technique was employed to transform the obtained point cloud model into a three-dimensional one with an uninterrupted surface.(2)Incremental octree map storage module.A 3D real-time reconstruction system of high precision can be utilized to reconstruct small-scale space,yet it is not sufficient for the 3D real-time reconstruction of a large outdoor environment.The existing point cloud model display map requires a vast memory space,making it unsuitable for real-time 3D reconstruction of large-scale environments.The existing octree map saving method can realize the octree map saving of different precision levels,but it can not be incremental storage,and expand the representation range of the map.In this paper,an incremental octree map preservation method is proposed,which uses the memory double buffering technology to realize the incremental construction of octree map,expands the representation range of Octree map,and realizes the large-scale real-time 3D reconstruction.(3)Simulation virtual interactive verification based on three-dimensional reconstruction scene.In this paper,the 1:1 model of overhead transmission line installed in the laboratory is used to carry out experiments,and the point cloud observed by the depth camera is matched by feature matching to realize the dense reconstruction of the working environment of the power inspection robot,and the Gazebo simulator in the ROS operating system is used to simulate the gravity field and other environments to verify the operation of the power inspection robot and verify the feasibility of the maintenance action planning of the overhead transmission line operation robot.By importing the three-dimensional model of the reconstructed transmission line and the URDF model of the inspection robot into the Gazebo simulator,the instructions are used to control the movement of the robot’s working arm to check whether the robot has problems such as model interference and unreasonable maintenance action.This paper uses the overhead transmission line 1 carried by the laboratory:1.Model experiments are carried out,and large-scale 3D real-time reconstruction is proposed based on RGB-D sensor and IMU information fusion.ORB-SLAM algorithm is used to reconstruct the map through feature matching of the point cloud observed by the depth camera,and then Octo Map is used to convert the real-time octree map.Octree’s 3D reconstruction map can be extended to address the challenge of large-scale 3D reconstruction.The three-dimensional reconstruction method of the work scene of the line patrol robot proposed in this paper realizes the function of large-scale reconstruction of the operating environment of the overhead transmission line,and establishes a simulation control platform for the feasibility verification of the maintenance operation,which effectively improves the intelligent level of the inspection robot for the overhead transmission line. |