Deep Learning-Based Classification of Weld Surface Defects

Zhu, Hui; Ge, Weimin; Liu, Zhenzhong

doi:10.3390/app9163312

Cited by 49 publications

(23 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, there have been publications devoted to the use of deep learning for automatic object recognition in materials science and related fields. For example, a number of studies were aimed at searching for defects in metals [ 12 , 13 , 14 , 15 , 16 ] including images of atomically resolved scanning transmission electron microscopy [ 17 ], classification of objects in scanning electron microscope images [ 18 ], and determining bubbles sizes in thermophysical processes [ 19 ].…”

Section: Introductionmentioning

confidence: 99%

Nanoparticle Recognition on Scanning Probe Microscopy Images Using Computer Vision and Deep Learning

et al. 2020

View full text Add to dashboard Cite

Identifying, counting and measuring particles is an important component of many research studies. Images with particles are usually processed by hand using a software ruler. Automated processing, based on conventional image processing methods (edge detection, segmentation, etc.) are not universal, can only be used on good-quality images and need to set a number of parameters empirically. In this paper, we present results from the application of deep learning to automated recognition of metal nanoparticles deposited on highly oriented pyrolytic graphite on images obtained by scanning tunneling microscopy (STM). We used the Cascade Mask-RCNN neural network. Training was performed on a dataset containing 23 STM images with 5157 nanoparticles. Three images containing 695 nanoparticles were used for verification. As a result, the trained neural network recognized nanoparticles in the verification set with 0.93 precision and 0.78 recall. Predicted contour refining with 2D Gaussian function was a proposed option. The accuracies for mean particle size calculated from predicted contours compared with ground truth were in the range of 0.87–0.99. The results were compared with outcomes from other generally available software, based on conventional image processing methods. The advantages of deep learning methods for automatic particle recognition were clearly demonstrated. We developed a free open-access web service “ParticlesNN” based on the trained neural network, which can be used by any researcher in the world.

show abstract

Section: Introductionmentioning

confidence: 99%

Nanoparticle Recognition on Scanning Probe Microscopy Images Using Computer Vision and Deep Learning

et al. 2020

View full text Add to dashboard Cite

show abstract

“…The results were experimental on our metal AM parts quality dataset. The accuracy obtained by histogram of oriented gradients (HOG) + SVM [42] was 79.6%, while the accuracy of 89.3% was achieved using Liu's CNN model [36]. Our approach had an accuracy of 92.1%.…”

Section: Performance Evaluationmentioning

confidence: 96%

“…Besides, CNN was adopted to link experimental microstructure with ionic conductivity for yttria-stabilized zirconia samples [32]. The CNN models have been applied in surface detection in bearing rollers, aluminum parts, and steel plates [33][34][35][36][37]. It was found out that CNN-based methods had better and more robust performance compared to the SVM classifiers.…”

Section: Introductionmentioning

confidence: 99%

Metal Additive Manufacturing Parts Inspection Using Convolutional Neural Network

Cui

Zhang

et al. 2020

Applied Sciences

View full text Add to dashboard Cite

Metal additive manufacturing (AM) is gaining increasing attention from academia and industry due to its unique advantages compared to the traditional manufacturing process. Parts quality inspection is playing a crucial role in the AM industry, which can be adopted for product improvement. However, the traditional inspection process has relied on manual recognition, which could suffer from low efficiency and potential bias. This study presented a convolutional neural network (CNN) approach toward robust AM quality inspection, such as good quality, crack, gas porosity, and lack of fusion. To obtain the appropriate model, experiments were performed on a series of architectures. Moreover, data augmentation was adopted to deal with data scarcity. L2 regularization (weight decay) and dropout were applied to avoid overfitting. The impact of each strategy was evaluated. The final CNN model achieved an accuracy of 92.1%, and it took 8.01 milliseconds to recognize one image. The CNN model presented here can help in automatic defect recognition in the AM industry.

show abstract

“…In order to verify the prediction accuracy of the proposed TL-MobileNet model, the TL-MobileNet model was compared with other models for welding defect detection. The other models include: back propagation (BP) [4], Knearest neighbors (KNN) [4], Extreme Learning Machine [18], Histogram of Oriented Gridients (HOG) [19], Convolutional neural networks (CNN) [31], artificial neural network (ANN) [11], support vector machines with principal component analysis (PCA-SVM) [32] and Extreme learning machine [17]). Table 6 represents the comparisons between the accuracy of the proposed method and that of other researchers in the prediction of welding defects (the input image size 96×96).…”

Section: ) the Influence Of Image Size M On Prediction Accuracymentioning

confidence: 99%

A New Image Recognition and Classification Method Combining Transfer Learning Algorithm and MobileNet Model for Welding Defects

et al. 2020

View full text Add to dashboard Cite

Welding quality directly affects the welding structure's service performance and life. Hence, the effective monitoring welding defects is essential to ensure the quality of the weld structure. Owing to the non-uniformity of the shape, position and size of welding defects, it is a complicated task to analyze and evaluate the acquired welding defects images manually. Fortunately, deep learning has been successfully applied to image analysis and target recognition. However, the use of deep learning to identify welding defects is time-consuming and less accurate due to the lack of adequate training data samples, which easily cause redundancy into the classifier. In this situation, we proposed a new transfer learning model based on MobileNet as a welding defect feature extractor. By using the ImageNet dataset (non-welding defect data) to pre-train a MobileNet model, migrate the MobileNet model to the welding defects classification field. This article suggested a new TL-MobileNet structure by adding a new Full Connection layer (FC-128) and a Softmax classifier into a traditional model called MobileNet. The entire training process of TL-MobileNet model has been successfully optimized by the DropBlock technology and Global average pooling (GAP) method. They can effectively accelerate the convergence rate and improve the classification network generalization. By testing the proposed TL-MobileNet on the welding defects dataset, it turned out our model prediction accuracy has arrived at 97.69%. The experimental results show that in several aspects, TL-MobileNet have better performance than other transfer learning models and traditional neural network methods.

show abstract

Deep Learning-Based Classification of Weld Surface Defects

Cited by 49 publications

References 32 publications

Nanoparticle Recognition on Scanning Probe Microscopy Images Using Computer Vision and Deep Learning

Nanoparticle Recognition on Scanning Probe Microscopy Images Using Computer Vision and Deep Learning

Metal Additive Manufacturing Parts Inspection Using Convolutional Neural Network

A New Image Recognition and Classification Method Combining Transfer Learning Algorithm and MobileNet Model for Welding Defects

Contact Info

Product

Resources

About