Cost and time-efficient method for accurate and high-resolution inspection of steel girder bridges


According to a recent America’s Infrastructure report, 42% of bridges are at least 50 years old and 7.5% of the nation’s bridges are considered structurally deficient. To ensure the safety of in-service bridges, the Federal Highway Administration recommends all bridges be inspected and documented periodically. Currently, the primary inspection method is visual inspection using physical methods. Despite being effective, these methods are quite time-consuming, expensive, and involve risks. The use of three-dimensional laser scanners, particularly unmanned aerial vehicles-based laser scanners, reduces problems associated with physical observations. Laser scanner systems produce laser point clouds that encode the geometric properties of structures, and the acquired laser point clouds should be post-processed to gather meaningful data that are essential for bridge inspection and management. The extraction of element-level data from point cloud maps is a key step for both laser-based surface defect detection and as-built bridge modeling. Existing methods either only address simple blocky structures or require substantial human interventions in data interpretation and processing, which is cost-intensive and time-consuming. This is particularly the case for bridges with thin members like steel girder bridges, where there is currently no automated inspection process. Therefore, this invention is aimed at automating the recognition and extraction of the main structural elements from laser point clouds of steel girder bridges.

Technology Overview

Researchers at Northeastern invented a collection of algorithms that can automatically convert laser scanned data captured in situ from a steel girder bridge of any type into a finite element (FE) mesh assembly, which reflects the as-built conditions of the bridge. The FE mesh assembly is comprised of a suite of conformal all-hexahedron FE meshes, each corresponding to an individual structural member (e.g., a steel girder, a cross-frame member, a bolt, or a gusset plate) in the bridge. In the FE mesh assembly, the connectivity and alignment between the structural members are accurately retained so that the FE meshes of each pair of connected structural members are assured to abut each other but do not overlap with each other. As such, the FE mesh assembly can be imported directly into FE analysis software (e.g., Abaqus) and be analyzed without the need for any manual mesh editing. This technology is intended to provide an automated and cost-effective approach for creating as-built FE models for steel girder bridges, which is key to refined load rating and evaluation of such bridges. In addition, the proposed method is effective in addressing the varying data qualities of laser point clouds, in terms of both levels of accuracy and resolution. 


  • Cost and time-efficient 
  • Automatic documentation of in situ conditions
  • Automatic processing of laser point cloud data 
  • High resolution and accuracy 


  • Civil engineering consulting firm 
  • Agencies responsible for assessing and fixing bridges/public structures


  • Licensing
  • Partnership 
  • Development collaboration 
Patent Information:
For Information, Contact:
Myron Kassaraba
Director of Commercialization
Northeastern University
Jerome Hajjar
Yujie Yan
Finite element mesh generation
Laser point cloud
Unmanned aerial vehicle