Real-time Multimodal Sensing System for Additive Construction by Extrusion: Integrating thermal, depth and RGB data

Abstract

Additive construction by extrusion addresses challenges such as labor shortages and inconsistent quality in traditional construction methods by offering automation, precision, and reduced waste. However, it remains challenging for large scale implementation due to dynamic printing conditions and environmental influences. Real-time control can be greatly beneficial in addressing these variations, which relies on a comprehensive and accurate real-time monitoring system. In this paper, we describe our ongoing work on developing a real-time multimodal sensing system that integrates thermal, depth and RGB data. By combining these sensors, the system can capture surface temperature and geometric data of extruded material, providing detailed feedback on its strength, extrudability and buildability. It provides new insights and information for quality monitoring systems in additive construction by extrusion, achieving a better balance between detailed detection and real-time monitoring. The methodology includes RGB-thermal image alignment, automated sample detection, and structured light-based depth sensing to refine geometric and temperature data. This system lays the groundwork for future advancements in predictive analysis and dynamic real-time printing control.