SBIR/STTR Award attributes
FTL's “Structural Damage And Repair Inferencing Tool (SDARIT)” leverages FTL’s 2D and 3D computer vision and machine learning pipeline. It provides an automated pier repair planning tool that accepts inputs from remote sensing systems and is capable of identification, volume approximation, location, and tabular summation of defects, all approximated from conventional construction practices. This system addresses an immediate Navy need for Port Damage Repair (PDR), which commonly includes gathering of 3D point-clouds. Currently, all key details of the scanned data are interpreted and entered through a process that is both manual and laborious. FTL proposes that this bottleneck can be alleviated by automatically converting structural 3D scan data into actionable Battle Damage Assessments (BDAs). SDARIT combines classically-coded algorithmic point-cloud dissection and pattern manipulation with neural network damage recognition. It does not rely on libraries of pre-modelled mechanical components, but address the inherent variability and lack of uniformity, precision, and accuracy in port structures. SDARIT will deliver volumetric construction details on both existing and missing components, while differentiating design variability from gross defects and battle damage.
