Point cloud technologies for smart cities: acquisition, processing, and applications

Abstract

Recent progress in LiDAR, UAV, and photogrammetric systems has made spatial data collection faster and more accessible. These tools enable the acquisition of detailed point clouds that form the foundation for many smart city applications. Efficient processing of these datasets is now a practical necessity, especially for everyday tasks such as monitoring roads and bridges, managing traffic, or building 3D city models used in digital twins. This paper reviews both classical and deep learning-based processing methods, data acquisition techniques, and multi-sensor integration strategies. Furthermore, the paper highlights applications beyond infrastructure, such as environmental monitoring of green areas and the analysis of pedestrian and bicycle networks. Despite the significant progress achieved in recent years, several open challenges remain. Among the most important are the need for standardized data formats, improved computational efficiency, and robust fusion of heterogeneous sensor data. Overcoming these difficulties is key to ensuring that digital twins and AI-based analysis become useful tools in practical urban management. Ultimately, continued progress in this field can make a meaningful contribution to the development of smarter and more sustainable cities.