Detecting defects in 3D printing has become crucial as additive manufacturing gains traction in key industries like aerospace, automotive, and healthcare. This paper offers a thorough review of the methods used for defect detection in 3D printing, highlighting image processing, machine vision, and the integration of deep learning techniques. It contrasts traditional methods, which depend on manual feature extraction and classification algorithms, with modern deep learning approaches that automate feature extraction and classification in a unified process. Additionally, the paper compares full-reference methods— where defects are detected by comparing printed parts against ideal reference models—with no-reference methods that identify anomalies without predefined models. The review also explores real-time monitoring systems that allow for early defect detection during printing, reducing production failures and material waste. Future developments are anticipated to focus on autonomous feedback mechanisms, fostering innovation in defect prevention and improving the sustainability of 3D printing processes.

Keywords