Automated Measurement Extraction from Laser Scanned Point Clouds toSupport Bridge Inspection

摘要

Dense point clouds collected by laser scanners enable bridge inspectors to reconstruct 3D as-builtmodels and take measurements on 3D models. That virtual inspection approach saves time duringfield surveying, and yet it still requires inspectors to manually identify geometric features requiredby a measurement, select data patches associated with those features, execute proper dataprocessing and feature extraction algorithms on those data patches, and calculate measurementresults based on extracted features. Those manual operations not only result in additional dataprocessing and interpretation time, but also requires users to understand technical characteristics ofthe scanners and data processing algorithms for extracting measurements with high accuracies.There is a need for automated support for virtual inspection/measurement extraction. This paperdescribes an approach providing automated support for that 3D measurement query and extraction.