Unsupervised Welding Defect Detection Using Audio And Video

TL;DR

This paper presents an unsupervised deep learning approach using audio and video data to detect welding defects in real-time, significantly improving accuracy across multiple defect types.

Contribution

It introduces a multi-modal, unsupervised deep learning method for real-time welding defect detection using audio and video data, trained on a large diverse dataset.

Findings

Achieved an average AUC of 0.92 across 11 defect types

Multi-modal approach outperforms single modality detection

Demonstrated real-time defect detection feasibility

Abstract

In this work we explore the application of AI to robotic welding. Robotic welding is a widely used technology in many industries, but robots currently do not have the capability to detect welding defects which get introduced due to various reasons in the welding process. We describe how deep-learning methods can be applied to detect weld defects in real-time by recording the welding process with microphones and a camera. Our findings are based on a large database with more than 4000 welding samples we collected which covers different weld types, materials and various defect categories. All deep learning models are trained in an unsupervised fashion because the space of possible defects is large and the defects in our data may contain biases. We demonstrate that a reliable real-time detection of most categories of weld defects is feasible both from audio and video, with improvements achieved by combining both modalities. Specifically, the multi-modal approach achieves an average Area-under-ROC-Curve (AUC) of 0.92 over all eleven defect types in our data. We conclude the paper with an analysis of the results by defect type and a discussion of future work.