Next-Generation Multispectral and Multimodal Imaging Architectures Enabled by Edge Processing
Luke Spinosa, EOPTIC
Abstract
Recent advances in edge computing have enabled a new generation of intelligent imaging systems capable of processing complex multimodal data directly on the camera. Leveraging compact, high-efficiency graphics processing units such as those developed by NVIDIA, modern electro-optical platforms can perform real-time fusion of visible, infrared, polarization, and event-based imagery to extract information traditionally inaccessible to standard electro-optical and infrared systems.
Eoptic has applied these capabilities across a wide range of domains, including explosive ordnance detection, counter-unmanned aircraft systems, space payloads, and industrial metrology. These efforts demonstrate how edge-enabled processing can fully utilize multispectral and multimodal data to improve contrast, material discrimination, and overall decision confidence. By combining physics-based spectral design with advanced processing, these systems transform imaging architectures from passive data collectors into intelligent analytic platforms. This scalable approach makes high-dimensional data more accessible and actionable across defense, space, and advanced manufacturing applications.
Bio
Luke Spinosa is an imaging systems engineer specializing in multimodal and multispectral sensing for defense and industrial applications. He holds a Bachelor of Science in Imaging Science from the Rochester Institute of Technology.
At Eoptic, Inc. in Rochester, New York, Luke has contributed to the development of advanced electro-optical systems that combine high-performance imaging hardware with real-time edge processing. His work spans synthetic data modeling, optical design, and artificial intelligence-based characterization, enabling fusion of multiple imaging modalities directly at the edge to provide rapid situational awareness.
As a principal investigator on multiple programs, Luke leads the design of next-generation imaging solutions that integrate multispectral UV-LWIR, polarization, and event-based sensing into compact, edge-enabled platforms. By strategically selecting spectral bands and sensing modalities informed by material and atmospheric physics, these systems extract features and distinctions that conventional electro-optical and infrared systems often overlook. They support missions across counter-unmanned aircraft systems, space-based payloads, autonomous vessels, and precision metrology
Parking and location
The talk will be held at UR River Campus, Goergen 108. Parking is available in the lot across the street in Intercampus Drive Lot, and is free for talk attendees (no pass needed).
Pre-talk dinner
A pre-talk dinner will be held at 5:15 pm at The King and I (1455 E Henrietta Rd, Rochester, NY 14623). If you would like to attend, please contact the house committee by email to make reservations. Email: house@opticarochester.org
