Welding defects classification through a Convolutional Neural Network. (January 2023)

• Record Type: Journal Article
• Authors: Perri, Stefania; Spagnolo, Fanny; Frustaci, Fabio; Corsonello, Pasquale
• Is Part Of: Manufacturing letters. Volume 35(2023)
• Page Start: 29
• Page End: 32
• Publication Date: 2023-01
• Subjects: Welding defect -- Convolutional Neural Network -- Dataset; Manufacturing industries -- Periodicals; Production engineering -- Periodicals; Manufacturing industries; Periodicals; 670
• DOI: 10.1016/j.mfglet.2022.11.006
• Languages: English
• ISSNs: 2213-8463

Abstract

Abstract: This letter presents a Convolutional Neural Network (CNN), named WelDeNet, customized to classify welding defects, such as lack of penetration (LP), cracks (CR), porosity (PO) and no defect (ND), by inspecting digitalized radiographic images. A new dataset that collects 24, 407 images representing welding defects is also presented. WelDeNet consists of 14 cascaded convolutional layers and achieves a test accuracy of 99.5 %. When hardware implemented within the Raspberry Pi 3B + board, WelDeNet exhibits an inference time of only 134 ms, with CPU and memory utilizations of just 51 % and 47 MB, thus offering a promising solution easy-to-integrate in a real industrial environment.