Characterizing Point Cloud Data Density for Spatial Change-Based Maintenance Planning of Civil Infrastructure Systems

Understanding the impacts of point-cloud data density on the reliability of change detection calculations is crucial for maintenance planning of civil infrastructure systems. Automated spatial change detection of structures (e.g., deformations of bridges) based on laser scanning data can help engineers assess structural conditions and prioritize maintenance activities. Collecting point cloud data with high accuracy and level of detail is time-consuming and difficult due to complexed environments on construction sites. An efficient method that evaluates the impacts of data quality on change detection is thus important for inspectors. Inspectors could collect point clouds with relatively low data densities to save time for data collection and data processing without compromising domain requirements of change detection in engineering projects. As a step toward comprehending 3D data quality on spatial change analysis of structures, this paper examines an automated 3D data quality checking method to quantify the impacts of point cloud data density on the reliability of spatial change detection of civil infrastructures. The authors designed five comparisons for supporting the impact analysis of data density. In the five comparisons, the authors used the same reference point cloud and five point clouds with different data densities. The authors sub sampled the point clouds to make them having lower, similar, and higher data densities than those of the reference point cloud. Based on the change detection results, the authors found that when the data density of the point cloud is similar as that of the reference point cloud, the change detection method detects most of the changed areas. When data density of the compared point cloud is much higher than that of the reference point cloud, the area of the detected changes increases a little but it takes significantly more time to collect and process the data. The future research will further characterize the impacts of data noise, accuracy, and data completeness on the reliability of change analysis of structures of various geometric complexities.