A surface cracks detection method for nuclear fuel pellets using an improved fully convolutional network. Issue 5 (4th May 2022)

• Record Type: Journal Article
• Authors: Liu, Xiaoqiang; Peng, Wenbin; Zhang, Bin; Wang, Xinmiao; Zhang, Xiaofang; Zhang, Wenjie; Miao, Yanjie; Wu, Ge; Zhang, Chaoying
• Is Part Of: Journal of nuclear science and technology. Volume 59:Issue 5(2022)
• Page Start: 555
• Page End: 563
• Publication Date: 2022-05-04
• Subjects: Fully convolutional network -- deep learning -- crack detection -- nuclear fuel pellet; Nuclear engineering -- Periodicals; Nuclear physics -- Periodicals; Nuclear energy -- Periodicals; 621.4805
• DOI: 10.1080/00223131.2021.1987347
• Languages: English
• ISSNs: 0022-3131

Abstract

ABSTRACT: Surface cracks, one of the main quality defects in nuclear fuel pellets, are usually subtle and hidden in complex backgrounds, making them difficult to detect accurately and threatening reactor safety. In order to achieve accurate cracks detection, a fuel pellet crack detection (FPCD) network is proposed based on the fully convolutional network with the VGG16 network structure as the encoder's backbone (FCN-VGG16). The new architecture modifies the combination mode of convolution layers and pooling layers, and optimizes the size and number of convolution kernels. And a large weight is added for crack samples in the loss function to solve the class imbalance problem. Subsequently, a self-acquired dataset containing 2240 pellet surface crack images is used for training and performance evaluation of networks. Experiments show that the proposed method achieved an F1-score of 88.89% and an intersection over union (IoU) of 80.30%, obtaining better detection results for complex background crack images.