The US Bureau of Reclamation is seeking to learn about existing, new, or emerging data processing and feature recognition solutions for crack mapping within photogrammetric orthomosaic data. A system should be able to process data sets to detect cracks at a resolution of 1/8” wide or better on large water control structures. Specific needs within this project are better methods to map crack locations, use of temporal data to monitor crack growth over time as a metric to determine structural health, and optimization of collected datasets to improve automated crack mapping. The accuracy of the automated system should be better than 90%.
The US Bureau of Reclamation has the capability to collect photogrammetric data from a variety of platforms (UAV/UAS, ground-based photos, underwater ROV, etc.). These data can then be used to develop high resolution 2D/3D renderings of the structures or individual features within each structure. These renderings are used by engineers to identify, quantify, and monitor damage and deterioration on the structures. The current procedure is time-intensive and takes time to receive results.
Possible Solution Areas
- Feature recognition software
- Machine learning
- Data processing and analysis software
Desired Outcome of the Solution
An ideal solution would be an automated system that can detect and map defects at Reclamation structures and can monitor the change in the cracks over time.
Field of Use and Intended Applications
Increase the efficiency of structural integrity analysis of dams.
Related Tech Needs
Photo Credit: Graphist and Operator Printing machines solvent/eco-solvent from Pixabay