Research On SLAM Method Of Indoor Scene Based On RGB-D

In recent years,with the rapid development of computer technology,SLAM technology has been widely used in the field of mobile robots,unmanned aerial vehicles,pilotless automobile,visual medical,AR/VR,wearable devices,and so on.The SLAM method based on vision has become a hot research topic at home and abroad,with the graph optimization problem in sparse matrix found in recent years,the SLAM method based on graph optimization in a large-scale scene has been widely used.This paper uses the ASUS Xtion Pro live depth camera as the sensor,the paper proposes a 3D SLAM method based on BoVW model,the SLAM method proposed in this paper,and the efficiency and robustness of the proposed method is proved by the experimental results of the SLAM method.Firstly,the basic principle and method of vision based SLAM are introduced.Description of the SLAM problem,the SLAM analysis of several classical methods,analysis and comparison of the advantages and disadvantages of several classic feature detection,feature extraction based on visual SLAM on request in the ORB feature extraction algorithm proposes an adaptive extraction method of region segmentation based on ORB features.And in the feature matching,using random sample consensus(RANSAC)algorithm and K nearest neighbor algorithm to eliminate false matching,effectively reduce the false matching points and improve the matching accuracy and speed.In the point cloud data fusion algorithm,the ICP algorithm is used to solve the pose of the camera by SVD method.Secondly,in the closed-loop detection method,introduces the function and method of Closed-loop Detection,problems and difficulties and closed-loop detection,the closed-loop detection method based on BoVW model,introduces the method to create a visual dictionary,compared with the traditional shortcomings of K-Means clustering algorithm,and proposed an improved K-Means algorithm,effectively solve the K-Means algorithm depends on the initial cluster center,easy tofall into local optimum,improve the accuracy of closed-loop detection.Finally,the paper designs an indoor scene SLAM framework based on RGB-D,and applies the improved algorithm to the SLAM method.