Abstract
Welding radiographic image analysis (WRIA) is a key technology for welding automated non-destructive testing. Although there already exist some valuable surveys on WRIA, they do not provide a systematic overview of the challenges faced by WRIA and lack a careful distinction and comparison of the core feature techniques in WRIA. With the rapid development of the WRIA area, it is both urgent and challenging to comprehensively review the relevant studies. Therefore, this paper provides an extensive review of 164 papers published in the recent quarter century (1997–2021) ever since its first coined. Three key aspects, namely the challenges faced by WRIA, the evolutionary paths involved in the technology, and the specific application tasks are further discussed in detail. At last, potential future perspectives for WRIA are explored in terms of problem setup, technical improvement, and humanistic care, so as to provide useful insights to both academic researchers and industrial practitioners.
Original language | English |
---|---|
Article number | 112821 |
Number of pages | 19 |
Journal | Measurement |
Volume | 214 |
DOIs | |
Publication status | Published - 15 Jun 2023 |
Keywords
- Deep learning
- Image analysis
- Non-destructive testing
- Radiographic image
- Welding
ASJC Scopus subject areas
- Instrumentation
- Electrical and Electronic Engineering