The Research Of Robot Simultaneous Localization And Mapping Based On RGBD Camera

With the development of robotics and artificial intelligence,intelligent mobile robot has been applied widely in many fields.And Simultaneous Localization and Mapping becomes a key technology of intelligent mobile robot.How to build a real-time and efficent Simultaneous Localization and Mapping system becomes a focus of research.There are many sensors that can be applied to Simultaneous Localisation and Mapping system,such as laser rader,odometry and camera.Simultaneous Localization and Mapping system based on camera stands out among many methods and becomes a mainstream approach,because camera is cheap and can offer rich information,and Simultaneous Localisation and Mapping system based on camera developed rapidly after that.The current Simultaneous Localization and Mapping system has perception aliasing problem and low recall in indoor environment.In oder to get rid of these problems,This paper implements a mobile robot Simultaneous Localisation and Mapping system based on RGBD camera.The thesis mainly covers the following aspects:1.In this paper,the basic theory of mobile robot Simultaneous Localisation and Mapping is reviewed.Then,the difficulty of mobile robot Simultaneous Localisation and Mapping is discussed.At the same time,this paper presents a thorough study of the basic theory and method of mobile robot loop closure detection,mainly analyzes the shortcoming of bag of words method in loop closure detection.2.This paper designed a visual odometry based on RGBD camera.RGBD camera is easy to use and can offer rich environment information.It directly gives the depth of object,so this paper use RGBD camera to build Simultaneous Localisation and Mapping system.RGB picture can be restored to 3D by RGBD camera,and relative pose is calculated by iterative algorithm.3.In order to overcome the weakness of bag of words method in loop closure detection.This paper presents a method of loop closure detection based on line segment and endpoint,describes the detailed procedure of our method,achieves the mobile robot loop closure detection algorithm based on line segment and endpoint.Experiments show that our algorithm can reduce the influence of perception aliasing,and performs good at indoor and outdoor environment.Our method is tested on NC,CC,L6 I and L6 O datasets.It's recall at 100% precision reaches 57.14%,67.59%,80.45% and 69.65%.4.This paper design a Simultaneous Localisation and Mapping system based on a RGBD camera Visual Odometry and a g2 o back-end optimization,The proposed loop closure detection method is applied to Simultaneous Localisation and Mapping system.At last,a mobile robot Simultaneous Localisation and Mapping system based on RGBD camera was achieved.the system experiment shows that our system can get a good environment mapping.