Applied AI in Defect Detection for Additive Manufacturing: Current Literature, Metrics, Datasets, and Open Challenges

Abstract

Defect detection in additive manufacturing refers to the evaluation of collected industrial images and the identification of parts that cause anomalies to optimize decision-making in an industrial production context. The advent of the Internet of Things and the widespread installation of electronic sensors, such as image sensors in industrial production lines, have expanded the processing capabilities of analytics tools. By extracting visual information via convolutional operations, deep learning-based algorithms have recently overcome drawbacks of traditional machine learning methods. This paper provides an analysis of contemporary defect detection techniques based on deep learning. Existing methods for defect detection algorithms in additive manufacturing are discussed. In terms of potential research to improve defect detection in additive manufacturing, the difficulties and emerging trends related to defect detection through deep learning are also outlined.