Lidar And Binocular Camera Fusion Inspection Robot Real Time Position And Mapping

In order to ensure the safe and reliable use of electric energy,substations need to operate safely and stably.Regular inspections are essential to maintain the safety and reliability of substations.Intelligent substation inspection robots are the main research direction of substation inspection technology in the future,and the real-time positioning and mapping of inspection robots are the basis for the research of inspection robots.When the inspection robot is working,it is equipped with different types of sensors according to the specific environment,and obtains the perception data and its own posture in the environment.A single sensor will have cumulative errors in the positioning of a mobile platform,and multi-sensor data fusion can overcome the cumulative errors caused by a single sensor,and it is the current mainstream solution to the problem of mobile platform positioning.In this thesis,aiming at the problem of autonomous positioning of inspection robots in the outdoor environment of substations,the real-time positioning and mapping(SLAM)method is adopted,and the front-end of real-time positioning and mapping is designed by multi-sensor fusion.Compared with commonly used positioning sensors,binocular cameras and lidar are selected as the research objects,and the real-time positioning and mapping of 3D lidar substations are experimented and analyzed.After studying the extended Kalman filter algorithm,combined with the multi-innovation identification theory,the multi-innovation extended Kalman filter algorithm is proposed.On this basis,a more reasonable improved multi-innovation extended Kalman filter algorithm is proposed based on the weight calculation idea of particle filter.Through this algorithm,the binocular camera and the lidar are fused.Finally,the real-time positioning and mapping of the fusion of lidar and binocular cameras are completed in the substation environment.Experimental results show that the improved multi-innovation extended Kalman filter algorithm proposed in this thesis can be used for data fusion of lidar and binocular cameras.In the substation environment,it can improve real-time positioning accuracy,and the established three-dimensional map is more similar to the actual environment.