Research On Indoor Navigation And Positioning Technology Based On BIM Technology

Since human beings spend most of their time in the indoor environment,services based on indoor navigation and positioning technology have a huge application market.Compared with indoor navigation and positioning technology based on UWB,Lidar,and Wi Fi as sensing modules,navigation and positioning technology based on vision sensors has the advantages of no need to install auxiliary equipment in the scene and the high popularity of visual sensors.Therefore,navigation and positioning technology based on vision sensors has attracted wide attention.On the other hand,with the increasingly mature building information modeling(BIM)technology,how to use the building information provided by BIM to optimize the existing indoor navigation and positioning technology and reduce the cost of navigation and positioning has become a research hotspot in recent years.Aiming at the three core problems of monocular visual navigation and positioning: camera orientation initialization,long-time tracking,and path planning,a corresponding solution is proposed and the proposed algorithm is verified by using data sets.The main work and innovation of the thesis are as follows:(1)An improved In Loc algorithm is proposed to initialize the camera pose.Because the data set in large-scale indoor space is reconstructed from RGB-D data captured at sparse locations,the traditional pose initialization algorithm often fails due to the sparsity of the database.Therefore,based on the study of the classical hierarchical visual localization method(In Loc),the thesis proposes a new step named "pose correction" to reduce the impact of database sparsity.Specifically,the number of matched features is increased by reconstructing the local features that can be observed at the estimated pose,to improve the accuracy of the final pose initialization.(2)Aiming at the problem of long-term tracking and positioning in indoor navigation,a tracking and positioning method based on BIM and a monocular camera is designed.The algorithm captures the image of the scene where the carrier is located by the monocular vision sensor,extracts the line features in the image by the canny algorithm,and realizes the active positioning of the camera carrier by combining the geometric information provided by BIM.Then,the location of the carrier at the next moment is predicted by the improved Kalman filter algorithm,and BIM provides geometric information for the location at the next moment according to this location,to realize tracking and positioning.(3)Aiming at the problems of slow convergence speed and easily falling into the local optimal solution of the ant colony algorithm in indoor navigation,an improved A-star-Ant colony algorithm was proposed.Firstly,the A-star algorithm is used to optimize the initial pheromone of the ant colony,to improve the efficiency of the initial path search of the ant colony.Then,the decision-making mechanism of ant wayfinding is improved to improve the convergence speed of the algorithm.Finally,the pheromone release mechanism of the ant is improved to avoid the algorithm falling into the local optimal solution.