Inline Defective Laser Weld Identification by Processing Thermal Image Sequences with Machine and Deep Learning Techniques (original) (raw)
Related papers
Diagnostics of welding process based on thermovision images using convolutional neural network
IOP Conference Series: Materials Science and Engineering, 2019
Arc welding used at automated workstations in large-scale production systems requires continuous assessment of welded joints quality. There are known classical methods and diagnostic systems based on the observation of welding current or arc voltage, while along with the development of deep learning methods, the interest in diagnostics by the use of images is increasing. The article presents results of research conducted for the process of joining two stainless steel materials (AISI 304 and AISI 316L) of various thicknesses by means of a fillet weld, aimed at developing a method of diagnosing the welding process using a convolutional neural network. Infrared images recorded using two thermovision cameras mounted on a test stand were used to diagnose the process. EWM Tetrix 351 welding machine operating in TIG technology was used as an executive element. Welds were made at different currents and arc welding voltages, as well as at different welding speeds, which had a direct impact o...
Welding Defect Detection with Deep Learning Architectures
Welding Principles and Application [Working Title], 2022
Welding automation is a fundamental process in manufacturing industries. Production lines integrate welding quality controls to reduce wastes and optimize the production chain. Early detection is fundamental as defects at any stage could determine the rejection of the entire product. In the last years, following the industry 4.0 paradigm, industrial automation lines have seen the introduction of modern technologies. Although the majority of the inspection systems still rely on traditional sensing and data processing, especially in the computer vision domain, some initiatives have been taken toward the employment of machine learning architectures. This chapter introduces deep neural networks in the context of welding defect detection, starting by analyzing common problems in the industrial applications of such technologies and discussing possible solutions in the specific case of quality checks in fuel injectors welding during the production stage.
Weld Defect Detection Based on Deep Learning Method
Welding is an important joining technology but the defects in welds wreck the quality of the product evidently. Due to the variety of weld defects' characteristics, weld defect detection is a complex task in industry. In this paper, we try to explore a possible solution for weld defect detection and a novel image-based approach is proposed using small X-ray image data sets. An image-processing based data augmentation approach and a WGAN based data augmentation approach are applied to deal with imbalanced image sets. Then we train two deep convolutional neural networks (CNNs) on the augmented image sets using feature-extraction based transfer learning techniques. The two trained CNNs are combined to classify defects through a multi-model ensemble framework, aiming at lower false detection rate. Both of the experiments on augmented images and real world defect images achieve satisfying accuracy, which substantiates the possibility that the proposed approach is promising for weld defect detection.
Scientific Reports
Laser welding is a key technology for many industrial applications. However, its online quality monitoring is an open issue due to the highly complex nature of the process. this work aims at enriching existing approaches in this field. We propose a method for real-time detection of process instabilities that can lead to defects. Hard X-ray radiography is used for the ground truth observations of the sub-surface events that are critical for the quality. A deep artificial neural network is applied to reveal the unique signatures of those events in wavelet spectrograms from the laser back-reflection and acoustic emission signals. The autonomous classification of the revealed signatures is tested on reallife data, while the real-time performance is reached by means of parallel computing. The confidence of the quality classification ranges between 71% and 99%, with a temporal resolution down to 2 ms and a computation time per classification task as low as 2 ms. This approach is a new paradigm in the digitization of industrial processes and can be exploited to provide feedbacks in a closed-loop quality control system.
Application of Deep Learning in Infrared Non-Destructive Testing
Proceedings of the 2018 International Conference on Quantitative InfraRed Thermography, 2018
Convolutional Neural Networks (CNN) is already known as strong tools in various fields particularly in image processing and computer vision. This paper aims to exploit the power of CNN for transform learning and utilizes it as an unsupervised feature extractor for analyzing defects in a steel specimen with float batten holes and Carbon Fiber Reinforced Plastic (CFRP) composite materials. A pre-trained CNN (ImageNet-VGG-f) has been used for extraction of the vectorized features along with a spectral angler mapper (SAM) to provide a score for defects presented in the image. Empirical results on two aforementioned datasets indicate a promising performance for application of heating and cooling based active thermography with a reasonable computational cost due to unsupervised nature of the algorithm.
Welding defect classification based on convolution neural network (CNN) and Gaussian kernel
2017 International Seminar on Intelligent Technology and Its Applications (ISITIA), 2017
Visual inspection process for weld defects still manually operated by human vision, so the result of the test still highly subjective. In this research, the visual inspection process will be done through image processing on the image sequence to make data accuracy more better. CNN as one of the image processing technique can determine the feature automatically which is suitable for this problem in order to classify the variation of each weld defect pattern. Classification using Convolution Neural Network (CNN) consist of two stages: extraction image using image convolution and image classification using neural network. Gaussian kernel used for blurring image, it helps the extraction of images without losing the main information from the original image, this filter also minimize the occurrence of interference or noise. Results of the classification used to get the category of weld defects with high accuracy as a variable of a weld inspection process whether the weld is pass the standard or not. The proposed system has obtained classification with validation accuracy of 95.83% for four different type of welding defect. The data input of this research is the result of images captured by a webcam.
IRJET- X-Ray Weld Defect Recognition Using Deep Learning Technique
IRJET, 2021
X-ray weld defect detection is a significant task in the industry which requires trained human experts and enough specialists for performing timely inspections. This paper proposes a deep learning based approach to identify different weld defects automatically from x-ray images. To employ this a dataset containing 250 x-ray images labelled for five different classes: gas pore, cluster porosity, crack, tungsten inclusion and no defect is developed. Then a Convolutional Neural Network model is designed and used for its training purpose. In addition we apply data augmentation technique to get better generalization performance with small dataset. Using this method we got a better test accuracy, which authenticates the possibility that the proposed approach is promising for weld defect recognition from x-ray images.
Deep Neural Networks for Defects Detection in Gas Metal Arc Welding
Applied Sciences
Welding is one of the most complex industrial processes because it is challenging to model, control, and inspect. In particular, the quality inspection process is critical because it is a complex and time-consuming activity. This research aims to propose a system of online inspection of the quality of the welded items with gas metal arc welding (GMAW) technology through the use of neural networks to speed up the inspection process. In particular, following experimental tests, the deviations of the welding parameters—such as current, voltage, and welding speed—from the Welding Procedure Specification was used to train a fully connected deep neural network, once labels have been obtained for each weld seam of a multi-pass welding procedure through non-destructive testing, which made it possible to find a correspondence between welding defects (e.g., porosity, lack of penetrations, etc.) and process parameters. The final results have shown an accuracy greater than 93% in defects classi...
Engineering Proceedings, 2021
Nowadays, automatic defect detection research by deep learning algorithms plays a crucial role, especially for non-destructive evaluation with infrared thermography. In deep learning research, the databases are the Achilles’ heel during the training in order to preserve optimized performance. In this work, we will present the infrared thermography sequences databases from the Universite Laval Multipolar Infrared Vision Infrarouge Multipolaire (MIVIM) research group for regular and irregular defect analysis in order to provide the best data collection resources for the pretraining of convolutional neural network and feature extraction analysis with future researchers and engineers. The databases will include infrared thermography sequences from regular and irregular defects of carbon fiber-reinforced polymer (CFRP), glass fiber-reinforced polymer (GFRP), plexiglass, aluminum, and steel, which could be available online for public use and research purposes.
Big Data and Cognitive Computing
In quality evaluation (QE) of the industrial production field, infrared thermography (IRT) is one of the most crucial techniques used for evaluating composite materials due to the properties of low cost, fast inspection of large surfaces, and safety. The application of deep neural networks tends to be a prominent direction in IRT Non-Destructive Testing (NDT). During the training of the neural network, the Achilles heel is the necessity of a large database. The collection of huge amounts of training data is the high expense task. In NDT with deep learning, synthetic data contributing to training in infrared thermography remains relatively unexplored. In this paper, synthetic data from the standard Finite Element Models are combined with experimental data to build repositories with Mask Region based Convolutional Neural Networks (Mask-RCNN) to strengthen the neural network, learning the essential features of objects of interest and achieving defect segmentation automatically. These r...