Catalog Item

S&T Project Number 20105 Final Report: Identifying Cracks in Concrete from Previously Collected UAS Data Using Deep Learning

Report summarizing automated concrete crack mapping using deep learning. Crack mapping concrete structures is a way to document and monitor cracks. In the past, crack mapping has been very labor intensive from data collection to documentation. The use of UAS and photogrammetry has allowed for faster and more comprehensive data collection and products including high-resolution orthoimages used to identify and document cracks. In addition, deep learning models can be used to automatically identify cracks from the orthoimages. This paper presents the process used to develop a deep learning model for automatic crack detection from data collected by UAS.
Download File Opens in new window
Generation Effort S&T Project 20105: Improving UAS-Derived Photogrammetric Data Analysis Accuracy and Confidence for High-Resolution Data Sets Using Artificial Intelligence and Machine Learning
Type Uploaded file(s)
File Type PDF
Publisher Bureau of Reclamation
Publication Date
Update Frequency not planned
Last Update Thursday, October 1st, 2020


The findings and conclusions of this work are those of the author(s) and do not necessarily represent the views of the Bureau of Reclamation.