Geometric Distortion Detection in Fused Filament Fabrication Using Augmented Reality and Computer Vision

Recent advancements in additive manufacturing, also known as 3D printing, have revolutionized the manufacturing industry by enabling the production of complex geometries and rapid prototyping. However, ensuring consistent quality in print parts remains a challenge. In this research, we address the quality monitoring issues in Fused Filament Fabrication (FFF) by proposing a system that combines computer vision and augmented reality techniques. The system measures geometric accuracy by comparing features of a 3D-printed part, captured using an ordinary camera, with the CAD model of the part that is projected in the same space. In the process, augmented reality is employed for camera calibration, pose estimation, and perspective projection, enabling accurate tracking and visualization of 3D-printed parts. Image processing techniques, including image segmentation and differencing, are applied to compare the virtual and real-world images, allowing the detection of geometric distortion. To evaluate the scalability and effectiveness of the proposed system, experiments were conducted by comparing virtual and real-world images of 3D-printed parts, captured from different camera positions and orientations. The results demonstrate the capability of the proposed solution to perform geometric distortion detection in close-to-real-time, which is an essential building block for in-situ control mechanisms in FFF applications.