Traditional Image Processing Techniques (IPT), used for automating the detection and classification of weld defects from radiography images, have their own limitations, which can be overcome by Deep Neural Networks (DNN). DNN produces considerably good results in fields which offer big dataset for it to train. DNN trained with small datasets by conventional methods produces less accurate results. This limits the use of DNN in many fields. This study focuses to overcome this limitation, by adopting transfer learning using Pre-trained deep convolutional neural networks. By this method, a weld defect radiographic image classifier, which can classify 14 types of weld defects, was constructed. 940 Image patches of weld defects were manually collected and labelled from GDXray database. Subsequently the features of this weld defect dataset were extracted using VGG16 and ResNet50 CNNs, both pre-trained on the ImageNet database. Then machine learning models such as Logistic Regression, Support Vector Machine (SVM) and Random Forest were trained on these extracted features. The Classifier based on SVM trained on features extracted by ResNet50 outperforms the other counter parts with an accuracy of 98%. In all these cases, transfer learning improves performance and reduces the training time and computational system requirements.INDEX TERMS Classifier, convolutional neural networks (CNN), radiographic images, transfer learning, weld defect detection.

No abstract

No abstract

Welding is a vital joining process; however, occurrences of weld defects often degrade the quality of the welded part. The risk of occurrence of a variety of defects has led to the development of advanced weld defects detection systems such as automated weld defects detection and classification. The present work is a novel approach that proposes and investigates a unique image-centered method based on a deep learning model trained by a small X-ray image dataset. A data augmentation method able to process images on the go was used to offset the limitation of the small X-ray dataset. Fine-tuned transfer learning techniques were used to train two convolutional neural network based architectures with VGG16 and ResNet50 as the base models for the augmented sets. Out of the networks we fine-tuned, VGG16 based model performed well with a relatively higher average accuracy of 90%. Even though the small dataset was spread across 15 different classes in an unbalanced way, the learning curves showed acceptable model generalization characteristics.

No abstract