Neuromorphic Imaging for In-Process Monitoring of Concrete 3D Printing

In this work for the first time, we examine the suitability and pros/cons of using event-based neuromorphic imagers for in-process monitoring of concrete additive manufacturing. There is interest in incorporating in-process monitoring into 3D concrete additive manufacturing systems for the purpose of measuring/controlling geometry as well as for detecting/preventing defects, to enable a more robust large-scale fabrication process. A substantial challenge with in-process monitoring is reducing the volume of data generated; observing large prints for extended periods of time can require prohibitive amounts of memory, inhibiting the ease of transfer, storage, and real-time analysis. Neuromorphic imagers are fundamentally different from conventional imagers in that they only detect light intensity changes at the individual pixels and report these changes as events. The result is a potential memory requirement reduction. This work represents a preliminary investigation into the use of event imagers for efficient and high-speed in-process monitoring of concrete additive manufacturing.