Abstract

The use of laser scanning in the realm of bridge assessment has been primarily limited to measuring large bridge dimensions. Given the scale of these measurements compared to standard accuracy metrics for LiDAR sensors, it follows that they may be captured with relatively small errors (<1%). This paper explores LiDAR’s capacity to (a) estimate smaller cross-section dimensions of an operating bridge, and (b) quantify the observed errors in terms of capacity calculations (as opposed to simple percent errors). To satisfy these objectives, sixteen LiDAR scans of an eleven-span steel girder bridge were completed under normal operating conditions. Various dimensional quantities were extracted from the data both directly and using standard plane-fitting approaches (Plane Fitting, Ransac). Results indicated that dimensions obtained from Plane Fitting resulted in flexural capacities 4%-7% less than those computed using the dimensions from the bridge plans. The Ransac method estimated errors within 7%-10%, while the dimensions obtained directly from the point cloud data resulted in capacity errors of 9%-13%. Due to common errors sources, all dimensions were conservatively estimated throughout this study. However, the observed distortion of elements due to fabrication stresses, showed to have the potential of overestimation of dimensions if planar assumptions are made.

摘要

在桥梁评估领域使用激光扫描主要限于测量大型桥梁尺寸。鉴于这些测量的规模与 LiDAR 传感器的标准精度指标相比，因此可以以相对较小的误差 (<1%) 捕获它们。本文探讨了 LiDAR 的能力，以 (a) 估计正在运行的桥梁的较小横截面尺寸，以及 (b) 在容量计算方面量化观察到的误差（而不是简单的百分比误差）。为了满足这些目标，在正常运行条件下完成了对一座十一跨钢梁桥的 16 次 LiDAR 扫描。直接和使用标准平面拟合方法（Plane Fitting，Ransac）从数据中提取各种维度量。结果表明，从平面拟合获得的尺寸导致弯曲承载力比使用桥梁平面图的尺寸计算的低 4%-7%。Ransac 方法估计误差在 7%-10% 以内，而直接从点云数据获得的维度导致容量误差为 9%-13%。由于常见的错误来源，在整个研究过程中，所有维度都是保守估计的。然而，观察到的由于制造应力引起的元件变形表明，如果做出平面假设，则可能会高估尺寸。由于常见的错误来源，在整个研究过程中，所有维度都是保守估计的。然而，观察到的由于制造应力引起的元件变形表明，如果做出平面假设，则可能会高估尺寸。由于常见的错误来源，在整个研究过程中，所有维度都是保守估计的。然而，观察到的由于制造应力引起的元件变形表明，如果做出平面假设，则可能会高估尺寸。