Enabling moisture monitoring and deformation analysis in wood during drying by 4D-CT imaging coupled with deep learning

The pivotal role of wood in promoting a sustainable future is underscored by its renewable properties, carbon sequestration capabilities, and versatility across applications such as construction, furniture, and energy production. Despite these advantages, the large energy consumption involved in the wood-drying process poses substantial challenges that must be overcome to realize its sustainability potential fully. To address these challenges, we have introduced an innovative method employing U-net deep learning image registration coupled with four-dimensional computed tomography (4D-CT). This approach enables three-dimensional (3D), time-resolved visualization of moisture migration and deformation on a pixel-level resolution during the wood-drying process. Our findings reveal the intricate details of the anisotropic nature of wood and its varying behaviors within and between annual rings. These insights lay a crucial foundation for advanced modeling, simulations, and predictions, ultimately facilitating the optimization of drying schedules and enhancing wood quality. This cutting-edge methodology represents an influential advancement in nondestructive wood analysis, which is promising for revolutionizing industrial practices and contributing to more sustainable wood utilization.