Color-space analytics for damage detection in 3D point clouds

Recent advances have enabled the use of modern remote sensing technologies to create high-resolution 3 D point clouds that capture the in-situ geometry of infrastructure systems. However, there is a need for new methods of analyzing and leveraging these complex data types. In this paper, the authors present an approach to quantifying textural deterioration and surficial damages manifested in point cloud data, such as corrosion or spall. This is achieved through geometric analysis of a nonlinear projection of the original color-space, and results in an algorithm that is generalizable to a variety of structural defects. The behavior of this algorithm is illustrated through both laboratory and field-scale experimental analysis, in which point cloud colorimetric differentials are identified automatically via the color-space damage detection algorithm and then compared with pixel-wise ground truth measurements. Overall measurement errors were on the order of 5-10%, with most error resulting from surface staining and surface reflectivity.