Ultrafast Dynamic Defect Inspection With Computational Neuromorphic Imaging

Computational neuromorphic imaging (CNI) is leveraged for ultrafast dynamic defect inspection, offering unparalleled capabilities for fast‐moving samples, complex illumination conditions, and real‐time edge detection. The CNI paradigm will enhance our set of inspection tools in challenging environments, paving the way for intelligent, precise, and low‐latency industrial diagnostics and expanding potential imaging applications. Abstract Inspecting surface defects is critical for ensuring quality in advanced manufacturing. However, current defect inspection techniques are largely limited to ideal imaging conditions and require vibration‐free environments when using traditional industrial cameras. These frame‐based cameras typically exhibit latency and a limited dynamic range, leading to vulnerabilities in production quality and restricting their efficiency in meeting increasing production demands. To address this challenge, a novel inspection paradigm is developed, using computational neuromorphic imaging (CNI) for ultrafast dynamic defect inspection. By leveraging the high temporal resolution and high dynamic range of event‐based sensors, CNI enables ultrafast, exceptional dynamic range yet cost‐effective defect inspection while minimizing perceptual and computational latency. Experimental results demonstrate event processing sampling times spanning 300‐fold in fast motion and dynamic ranges exceeding 10000‐fold under varying illumination. Additionally, event‐driven data facilitates direct edge detection and visualization of defects, which is critical for real‐time continuous diagnosis. Vibration is also utilized to enhance structural defect detection and improve robustness in practical settings. CNI offers unique advantages for industrial applications, enabling defect inspection across diverse products and complex conditions. This approach opens up a new route toward ultrafast industrial inspection in challenging environments, addressing the growing demands of mass real‐time inspection and intelligent diagnosis in precision manufacturing. Computational neuromorphic imaging (CNI) is leveraged for ultrafast dynamic defect inspection, offering unparalleled capabilities for fast-moving samples, complex illumination conditions, and real-time edge detection. The CNI paradigm will enhance our set of inspection tools in challenging environments, paving the way for intelligent, precise, and low-latency industrial diagnostics and expanding potential imaging applications. Abstract Inspecting surface defects is critical for ensuring quality in advanced manufacturing. However, current defect inspection techniques are largely limited to ideal imaging conditions and require vibration-free environments when using traditional industrial cameras. These frame-based cameras typically exhibit latency and a limited dynamic range, leading to vulnerabilities in production quality and restricting their efficiency in meeting increasing production demands. To address this challenge, a novel inspection paradigm is developed, using computational neuromorphic imaging (CNI) for ultrafast dynamic defect inspection. By leveraging the high temporal resolution and high dynamic range of event-based sensors, CNI enables ultrafast, exceptional dynamic range yet cost-effective defect inspection while minimizing perceptual and computational latency. Experimental results demonstrate event processing sampling times spanning 300-fold in fast motion and dynamic ranges exceeding 10000-fold under varying illumination. Additionally, event-driven data facilitates direct edge detection and visualization of defects, which is critical for real-time continuous diagnosis. Vibration is also utilized to enhance structural defect detection and improve robustness in practical settings. CNI offers unique advantages for industrial applications, enabling defect inspection across diverse products and complex conditions. This approach opens up a new route toward ultrafast industrial inspection in challenging environments, addressing the growing demands of mass real-time inspection and intelligent diagnosis in precision manufacturing. Advanced Science, Volume 12, Issue 44, November 27, 2025.