Design And Implementation Of Indoor Personnel Positioning System Based On Multi-Information

With the rapid advancement of science and technology,the demand for personnel positioning systems in society also rapidly increase.As the outdoor personnel positioning technology has matured,the indoor personnel positioning technology become the research direction of scientific research workers.The generation and development of indoor personnel positioning system are of great significance for the advancement of multi-sensor fusion technologies,information processing technologies,and computer-related technologies.On this basis,an indoor personnel positioning system based on multi-information that fused computer vision positioning technology and multi-sensor fusion technology has been designed and implemented in this thesis,which provide a theoretical and practical reference value for indoor personnel positioning technology to some extent.The indoor personnel positioning system based on multi-information has took the point cloud data containing multiple information collected by the sensor as the basis of the indoor personnel positioning system,applied the human target detection,tracking and positioning algorithm,and finally completed the positioning of the indoor personnel.For the traditional human target detection algorithm,the detection rate is low when the illumination changed and the shadows are blocked.This thesis used the point cloud data and fused the Histogram of Oriented Gradient(HOG)feature and Histogram of Oriented Depth(HOD)feature on the basis of the traditional human target detection algorithm,which effectively reduced the impact of the light changes and occlusion issues on the detection of human target.The traditional particle filter algorithm is easily affected by the occlusion factors in tracking the target of the human.In this thesis,color information and depth information are integrated based on the traditional particle filter algorithm in order to effectively decrease the impact of occlusion on human target tracking.The indoor personnel positioning algorithm first aligned the color image and the depth image of the human target tracking result,and then performed the conversion of different coordinate systems.Finally,the coordinate of the gravity center point containing the human target region is projected onto the ground,and the indoor personnel positioning requirements could be achieved.Through the comparison and analysis of experimental results,it is verified that the human target detection,tracking and positioning algorithm proposed in this thesis could effectively lower the impact of light changes and occlusion issues on human target detection,tracking and positioning.Indoor positioning accuracy of 7.2m in length and 4.2m in width could reach within ± 20cm.