Pavement Crack Detection using Deep Learning and High-Resolution Drone Imagery

Session XPO24-95

Lockheed Martin Aeronautics (LMA) has many large sites across the nation with thousands of acres of pavement which require year-round inspection and maintenance to ensure the safety of their employees and operations. To maximize resource efficiency, Lockheed Martin, in partnership with 1898 & Co., has developed and adopted a Deep Learning Model technology by using drone imagery in ESRI ArcGIS Pro to detect and identify potential pavement deficiencies. Combined with annual site drone surveys, this tool will allow LMA Facilities personnel to target their maintenance efforts and implement cost effective corrective actions more effectively. Additionally, using digital maps on company owned mobile devices, the Facilities team can quickly navigate to specific areas of interest. They can then perform ground inspections and collect additional data including severity rating, comments, and images of the deficiency. All data is uploaded and stored on the LMA GIS platform, allowing it to be mapped and shared across the organization for further analysis.