3D Pavement Pit Point Cloud Correction

Urbanization's pace makes urban road surface conditions essential for city operation efficiency and resident quality of life. This paper proposes three-dimensional point cloud correction techniques to detect and assess road surface diseases like potholes, which can impact road performance and safety. It introduces equipment for data collection, including the Orbbec Astra Mini and Intel RealSense D435, and addresses challenges in point cloud data processing. The paper presents a novel multi-scale denoising algorithm, KmeansPT-DBSCAN, and an improved plane fitting algorithm, MINI_BATCH-RANSAC, enhancing data quality and processing speed. These techniques support automated road disease detection and assessment, benefiting road management and urban infrastructure technology application.