Comparative Study of Deep Learning and Machine Learning Techniques for Corrosion and Cracks Detection in Nuclear Power Plants

The detection of corrosion and cracks in nuclear power plants is a critical task that requires accurate and efficient monitoring systems. Traditional inspection methods can be time-consuming and may not be able to detect defects in hard-to-reach areas. Machine learning and deep learning have shown promising results as replacements for conventional ways of detecting corrosion and cracks in nuclear power reactors in recent years. This paper compares the latest research on machine/deep learning techniques for corrosion and crack detection in nuclear power plants. It includes an overview of the different machine/deep learning algorithms that have been applied in this field. Furthermore, this paper also investigates the effect of different input features and transfer learning techniques on the accuracy of corrosion and crack detection models. Finally, a systematic review of publicly available datasets for corrosion and crack detection in nuclear power plants is presented.