The Application of Point Cloud Data Registration Algorithm Optimization in Smart City Infrastructure

Abstract

Point cloud data, as an effective carrier for storing higher-dimensional spatial information, plays an important role in promoting the realization of smart cities. However, variations in spatial information, collection methods, and densities among multi-source heterogeneous point cloud data significantly reduce the accuracy of modeling and perception. This paper focuses on core aspects of point cloud registration — feature extraction, improved ICP algorithms, deep learning-based registration models, and multi-source fusion methods — to enhance registration accuracy and robustness. This paper further explores the implementation of point cloud registration methods in areas such as urban traffic modeling, building and terrain renewal, facility monitoring, and urban planning. Precise matching algorithms provide fundamental algorithmic guarantees and data foundations for effectively enhancing the efficiency of multi-source data fusion and for urban spatial modeling, dynamic monitoring, intelligent scheduling, etc.