Research on Fully Automated and Semi-Automated Single-Building Model Production Based on Airborne LiDAR

3D real scene as the spatiotemporal digital foundation for the future digital world, functioning as a crucial component of a nation's spatial information infrastructure. Given the numerous drawbacks of traditional 3D grid models in photorealistic representation, the focus of research has shifted towards the rapid and efficient production of single-building models. This paper explores two methodologies, namely fully automated and semi-automated approaches, for the production of single-building models. Utilizing the LiDARPro processing system in conjunction with multi-perspective aerial imagery and airborne laser point cloud data, the study compares the outcomes in terms of processing efficiency, model intricacy, and accuracy. The findings indicate that both approaches enable the swift construction of building models. While the fully automated process demonstrates higher efficiency, the semi-automated process excels in capturing realism and achieving higher model accuracy. Consequently, a combined approach of the two methods proves to be effective in attaining automated single-building model production in practical applications.