Color-based registration of point cloud data by video camera for electromagnetic simulation

摘要

Radio propagation simulation is a valuable tool to predict the propagation channel characteristics for system design and service-cell planning. In the simulation, the accurate environment model is needed to improve the prediction accuracy of the site-specific channel. Recently, the ability of smartphones makes remarkable progress, and the point cloud data of the environment can be obtained by the video camera output of the smartphone. However, because of the memory size limitation of those devices, it is necessary to repeat the measurement and combine the several point cloud datasets to model the large-volume environment. In this paper, we proposed the color-based registration method by extending the fast descriptor algorithm. We applied our proposed algorithm to the point cloud data taken by the video camera of the smartphone, and proved that the calculation time was reduced by 70% in the office scenario. The work is expected to be utilized for the radio propagation simulation in actual environments.