Convolution neural network model with improved pooling strategy and feature selection for weld defect recognition

Abstract

Weld defect recognition plays an important role in the manufacturing process of large-scale equipment. Traditional methods generally include several serial steps, such as image preprocessing, region segmentation, feature extraction, and type recognition. The results of each step have significant impact on the accuracy of the final defect identification. The convolutional neural network (CNN) has strong pattern recognition ability, which can overcome the above problem. However, there are two problems: one is that the pooling strategy has poor dynamic adaptability, and the other is the insufficient feature selection ability. To overcome these problems, we propose a CNN-based weld-defect recognition method, which includes an improved pooling strategy and an enhanced feature selection method.According to the characteristics of the weld-defect image, an improved pooling strategy that considers the distribution of the pooling region and feature map is introduced. Additionally, in order to enhance the feature selection ability of the CNN, an enhanced feature selection method integrating the ReliefF algorithm with the CNN is proposed. A case study is presented for demonstrating the proposed techniques.The results show that the proposed method has higher accuracy than the traditional CNN method, and establish that the proposed CNN-based method is successfully applied for weld-defect recognition.

Automated defect detection in digital radiography of aerospace welds using deep learning

“…Due to the much larger test set sample size in this study (about 500 on average), the lower 95% confidence bound (dashed line) is close to the POD curves. Considering the patch size (512 × 512 pixels), the material was large in comparison to other weld data sets [23,35].…”

“…The authors also demonstrated that this conventional data augmentation has limited scope and additional data augmentation offers diminishing returns. Jiang et al [23] proposed a novel pooling strategy for CNNs to better represent dark and light defects like slag and tungsten inclusion, and classified defects into six categories: crack, lack of fusion, lack of penetration, slag inclusion, porosity and non-defect. They used a set of 3486 32 × 32 pixel images divided into the six categories for training.…”

“…Jiang et al [23] used a CNN with classification output for weld defect detection. Segmentation offers advantages over classification in NDE applications: Defects can be measured and counted, and their shape can be determined, which is important for defect acceptance as discussed in Section 1.1.…”

A PRISMA-driven systematic review of data mining methods used for defects detection and classification in the manufacturing industry

“…Most recently, for example, the article "Unsupervised Pre-Training of Imbalanced Data for Identification of Wafer Map Defect Patterns" written by Shon et al (2021), where the proposed model is based on convolution variational autoencoder (CVAE), achieved a 95.1% level of F-measure. In our analysis, we found out that more than 42% of composed primary studies implementing NN are focused on CNN, for example, Lin et al (2019), Wang et al (2018) 2020), Cerezci et al (2020) or Jiang et al (2021). R-CNN algorithm is also often used (more than 13%), for example, in Tabernik et al (2019), Shi et al (2020a), Liyun et al (2020), Zhao et al (2020a), Zhang & Shen (2021) or Zhao et al (2021).…”

An Improved DeepNN with Feature Ranking for Covid-19 Detection

“…They have used the deep neural network (DNN) and multilayer bi-directional long short-term memory (BiLSTM) with the attention mechanism to forecast overdue repayment behavior. Jiang et al [17] have proposed an improved feature selection approach by integrating the CNN with the Relief algorithm. Also, Kaddar et al in [7] have used the ANOVA technique to find the non-redundant representation in CNN by obtaining the feature maps with various neuron responses.…”