Research And Implementation Of Mobile Robot Mapping Technology Based On Vision And Laser

In the field of mobile robots,mapping is a key research and the premise for autonomous navigation.The core technology of mapping is Simultaneous Localization and Mapping(SLAM).It means a robot estimates its own motion and builds a map of the environment at the same time when it moves in an unknown environment.The selection and application of map construction methods for outdoor mobile robots is one of its design difficulties.SLAM can be divided into lidar based SLAM and camera based SLAM according to the sensor type.Lidar is difficult to be interfered by light and has a longer detection distance.The camera's information is more abundant,and it doesn't affected by the terrain.With the visual information,robot can get the 3D posture and mileage.Therefore,it is meaningful to combine the advantages of the two methods by using visual SLAM to provide auxiliary information for laser SLAM in the outdoor environment.This thesis aims at the mapping technology in outdoor environments.The main research and work are as follows:Firstly,the ORB-SLAM2 based on ORB(Oriented FAST and Rotated BRIEF,ORB)feature extraction,which consuming too more time.To improve the method,the thesis optimized the tracking thread by using the sparse direct method with fast calculation speed to process non-key frames,while retaining the key frame selection strategy and back-end optimization module of ORB-SLAM2.And then,the new method realized a real-time visual odometry.The new method can reduce the amount of calculation and improving the operation speed of the algorithm while basically guaranteeing the accuracy.Experimental results show that the new method improves the real-time performance of ORB-SLAM2.The second research of the thesis is the mapping technology based on mobile robots.The Hector method,Gmapping method and Cartographer method are compared to analyze the performance the algorithm by taking the map accuracy and the consumption of computing resources as indicators.Through the results of experiment,Cartographer was finally selected as the mobile robot's mapping algorithm.The improved visual odometer is used to provide Cartographer with an initial pose for reference.Finally,the algorithm is deployed on the mobile robot for indoor and outdoor mapping experiments.The restoration effect of the map on the actual scene verifies the effectiveness of the algorithm.Thus,a feasible solution for outdoor map construction of mobile robots is provided.