STRIKE: System for Training Recognition and Identification of Known Explosives

Explosive ordnance disposal (EOD) operations face significant challenges in rapidly and accurately identifying a vast array of ordnance types, particularly under conditions where physical damage, environmental degradation, or limited visual access obscure critical identifying features. Current identification methods rely heavily on EOD technicians using static reference materials, such as photographs and line drawings, which are time-consuming and prone to human error. The lack of automated, high-fidelity imaging solutions limits the ability to train artificial intelligence/machine learning (AI/ML) models for robust detection and classification of ordnance in complex and uncontrolled environments. Furthermore, existing photogrammetry systems capture only visible light data, making them incompatible with advanced sensors, such as those operating in shortwave infrared (SWIR) or longwave infrared (LWIR) bands, used in modern robotics and autonomous systems. These limitations highlight a critical need for a comprehensive, scalable solution that integrates multispectral imaging and advanced 3D modeling to enhance EOD safety and operational efficiency.To address this challenge, Eoptic proposes the System for Training Recognition and Identification of Known Explosives (STRIKE), a metrology tool used to gather spatial and spectral measurements of real-world ordnance an EOD technician may encounter. STRIKE leverages structured light stereoscopy and radiometric calibration to achieve precise 3D surface mapping with accurate spectral data, enabling simulations under varying conditions based on proven optical metrology and computational imaging techniques. It completes the process in 4-7 minutes, depending on ordnance size and resolution needed. The system integrates spectral data directly into point cloud generation, creating digital twins of ordnance objects with unparalleled spatial and spectral fidelity. This enriched dataset enables the development of robust AI/ML models for ordnance identification, that can then be optimized for use on a variety of platforms, including handheld devices, ground robots, and unmanned aerial systems (UAS). STRIKE also incorporates an automated calibration pipeline and modular design, ensuring adaptability to evolving imaging technologies and operational needs.