Classification Of Defect Types Research Articles

Radiometric Infrared Thermography of Solar Photovoltaic Systems: An Explainable Predictive Maintenance Approach for Remote Aerial Diagnostic Monitoring

Solar photovoltaic (SPV) arrays are crucial components of clean and sustainable energy infrastructure. However, SPV panels are susceptible to thermal degradation defects that can impact their performance, thereby necessitating timely and accurate fault detection to maintain optimal energy generation. The considered case study focuses on an intelligent fault detection and diagnosis (IFDD) system for the analysis of radiometric infrared thermography (IRT) of SPV arrays in a predictive maintenance setting, enabling remote inspection and diagnostic monitoring of the SPV power plant sites. The proposed IFDD system employs a custom-developed deep learning approach which relies on convolutional neural networks for effective multiclass classification of defect types. The diagnosis of SPV panels is a challenging task for issues such as IRT data scarcity, defect-patterns’ complexity, and low thermal image acquisition quality due to noise and calibration issues. Hence, this research carefully prepares a customized high-quality but severely imbalanced six-class thermographic radiometric dataset of SPV panels. With respect to previous approaches, numerical temperature values in floating-point are used to train and validate the predictive models. The trained models display high accuracy for efficient thermal anomaly diagnosis. Finally, to create a trust in the IFDD system, the process underlying the classification model is investigated with perceptive explainability, for portraying the most discriminant image features, and mathematical-structure-based interpretability, to achieve multiclass feature clustering.

Anomaly Detection and Inter-Sensor Transfer Learning on Smart Manufacturing Datasets.

Smart manufacturing systems are considered the next generation of manufacturing applications. One important goal of the smart manufacturing system is to rapidly detect and anticipate failures to reduce maintenance cost and minimize machine downtime. This often boils down to detecting anomalies within the sensor data acquired from the system which has different characteristics with respect to the operating point of the environment or machines, such as, the RPM of the motor. In this paper, we analyze four datasets from sensors deployed in manufacturing testbeds. We detect the level of defect for each sensor data leveraging deep learning techniques. We also evaluate the performance of several traditional and ML-based forecasting models for predicting the time series of sensor data. We show that careful selection of training data by aggregating multiple predictive RPM values is beneficial. Then, considering the sparse data from one kind of sensor, we perform transfer learning from a high data rate sensor to perform defect type classification. We release our manufacturing database corpus (4 datasets) and codes for anomaly detection and defect type classification for the community to build on it. Taken together, we show that predictive failure classification can be achieved, paving the way for predictive maintenance.

Intelligent Detection and Classification of Surface Defects on Cold-Rolled Galvanized Steel Strips Using a Data-Driven Faulty Model With Attention Mechanism

Abstract In the production of cold-rolled galvanized steel strips used for stamping car body parts, the in-situ and real-time defect detection is crucial for quality control, in which various types of defects inevitably occur. It is challenging to improve the accuracy of defect detection and classification by appropriate means to assist the manual screening process better. Defects under actual production conditions are often not prominent enough in defect characteristics, and there may be a significant similarity between different defect categories. To eliminate this weakness, we propose a data-driven deep learning approach named steel surface faulty detection attention net (SSFDANet) that uses images of the galvanized steel surfaces as input to identify whether the product is qualified and automatic classification of defect types instantaneously. This method can shorten product inspection time and improve the production line automation efficiency. In addition, the attention mechanism is utilized to enhance the performance of SSFDANet. Compared with the baseline ResNet, SSFDANet achieves a noticeable improvement in classification accuracy on test data. The well-trained model can successfully show an improved performance than the baseline models on the multiple types of faulty. Enhanced by SSFDANet with high classification accuracy, the defect rate of products is significantly reduced, and the production speed of the production line is significantly improved. Future prospective studies that are inspired by this article are also discussed.

レーザー散乱光の多方向偏光検出によるSiウエハ上微小欠陥の高精度凹凸分類

Wafer defect inspection is essential to control quality and yield of semiconductor device manufacturing. Wafer surface inspection tools with laser scattering method are employed for high-throughput and high-sensitivity inspection of defects with sizes of about several microns to sub-20 nm. In addition to detecting defects with various shapes and sizes, classification of defect types is required. Classification by multi-directional scattered light detection is known, which captures defect scattered light at multiple scatter direction and compares the intensities in different angles. We simulate concave-convex defect classification with multi-directional detection by analyzing polarized angular scattering distributions, obtained with discrete dipole approximation method, from defects that differ in shapes, sizes, and materials. As a result of classification performance evaluation with support vector machine classifier, when using only unpolarized light intensities, the maximum defect diameter that is accurately classified is limited to from 0.62 times as much as the illumination wavelength. By using both +45 and -45 deg linearly polarized light intensities, the limit is extended to from 1.47 times as much as the wavelength.

A Study on the Detection of Internal Defect Types for Duct Depth of Prestressed Concrete Structures Using Electromagnetic and Elastic Waves.

Prestressed concrete (PSC) is widely used for the construction of bridges. The collapse of several bridges with PSC has been reported, and insufficient grout and tendon corrosion were found inside the ducts of these bridges. Therefore, non-destructive testing (NDT) technology is important for identifying defects inside ducts in PSC structures. Electromagnetic (EM) waves have limited detection of internal defects in ducts due to strong reflections from the surface of the steel ducts. Spectral analysis of the existing impact echo (IE) method is limited to specific conditions. Moreover, the flexural mode in upper defects of ducts located at a shallow depth and delamination defects inside ducts are not considered. In this study, the applicability of the elastic wave of IE was analyzed, and multichannel analysis of surface, EM, and shear waves was employed to evaluate six types of PSC structures. A procedure using EM waves, IE, and principal component analysis (PCA) was proposed for a more accurate classification of defect types inside ducts. The proposed procedure was effective in classifying upper, internal, and delamination defects of ducts under 100 mm in thickness, and it could be utilized up to 200 mm in the case of duct defect limitations.

TireNet: A high recall rate method for practical application of tire defect type classification

Tire enterprises often use X-ray photographs to inspect tire defects, which is still in the stage of manual inspection. Existing methods have achieved a lot of success in tire defect detection, but have not yet reached the application level requirements. It is hard for unsupervised methods to handle the complex specs and patterns of tires, while the supervised methods suffered from a lack of enough data. In this paper, an end-to-end method (TireNet) is proposed for the practical application of automatic tire defect detection with X-ray images. Object detection models are introduced as the baselines, it is found that the recall rate (missed detection rate) cannot meet the application-level needs of enterprises, since tire defects are quite different from general objects. Inspired by the features of tire X-ray images, which are periodic, the Siamese network is used in the new model as part of the downstream classifier to capture the defective features. Additionally, 120,000 labeled tire images (including qualified tires) were set up and collected as datasets. The experimental results of this method reach a higher recall rate compared to Faster R-CNN, SSD, and YOLO in the labeled datasets. At present, the model has been successfully deployed to the quality inspection department of enterprises and partly replaced artificial jobs. The application results show that the missing rate of TireNet is 0.17%, which is lower than that of the artificial 2.4%. With the help of a deployed assistant detection system, the tire defect detection rate increased by more than 100%, and the tire refund rate decreased by approximately 20%.

Deeper and Mixed Supervision for Salient Object Detection in Automated Surface Inspection

In recent years, researches in the field of salient object detection have been widely made in many industrial visual inspection tasks. Automated surface inspection (ASI) can be regarded as one of the most challenging tasks in computer vision because of its high cost of data acquisition, serious imbalance of test samples, and high real-time requirement. Inspired by the requirements of industrial ASI and the methods of salient object detection (SOD), a task mode of defect type classification plus defect area segmentation and a novel deeper and mixed supervision network (DMS) architecture is proposed. The backbone network ResNeXt-101 was pretrained on ImageNet. Firstly, we extract five multiscale feature maps from backbone and concatenate them layer by layer. In addition, to obtain the classification prediction and saliency maps in one stage, the image-level and pixel-level ground truth is trained in a same side output network. Supervision signal is imposed on each side layer to realize deeper and mixed training for the network. Furthermore, the DMS network is equipped with residual refinement mechanism to refine the saliency maps of input images. We evaluate the DMS network on 4 open access ASI datasets and compare it with other 20 methods, which indicates that mixed supervision can significantly improve the accuracy of saliency segmentation. Experiment results show that the proposed method can achieve the state-of-the-art performance.

Classification of Defect Types in SLM Ti-6Al-V4 by X-ray Refraction Topography

Abstract Porosity in additively manufactured materials, such as laser powder bed fusion Ti-Al6-V4, can play an important role in their mechanical performance. Not only the total porosity but also the shape/morphology of the individual pores need to be considered. Therefore, it is necessary to determine the distributions of different defect types (especially fusing defects and keyhole pores) and their dependence on process parameters. We show that synchrotron X-ray refraction radiography allows analysis of large samples (up to several millimeters) without compromising the detectability of submicrometer defects. Correspondingly, a classification tool is introduced that is able to quantitatively distinguish defects such as keyhole pores and binding defects with a confidence level of 94 %, even when the shape cannot be discerned because of limited spatial resolution.

Deep Regression Neural Network for Industrial Surface Defect Detection

Industrial product surface defect detection is very important to guarantee high product quality and production efficiency. In this work, we propose a regression and classification based framework for generic industrial defect detection. Specifically, the framework consists of four modules: deep regression based detection model, pixel-level false positive reduction, connected component analysis and deep network for defect type classification. To train the detection model, we propose a high performance deep network structure and an algorithm to generate label data to capture the defect severity information from data annotation. We have tested the method on two public benchmark datasets, AigleRN and DAGM2007, and an in-house capacitor image dataset. The results have shown that our method can achieve the state-of-the-art performance in terms of detection accuracy and efficiency.

Effective automatic defect classification process based on CNN with stacking ensemble model for TFT-LCD panel

The classification of defect types during LCD panel production is very important because it is closely related to deciding whether a defect panel is restorable. But since defect areas are very small compared to the panel area, it is hard to classify defect types by images. Therefore, we need to eliminate the background pattern of the panel, but this is not an easy task because the brightness and saturation of the background varies, even in a single image. In this paper, we propose an indicator that can distinguish between defect and background area, which is robust to brightness change and minor noises. With this indicator, we got useful defect information and images with patterns eliminated to make a more efficient defect classifier. The convolutional neural network with stacked ensemble techniques played a great role in improving defect classification performance, when various information from image preprocessing was combined.

SMT defect classification by feature extraction region optimization and machine learning

In this paper, we propose a solder joint defect type classification method for automatic optical inspection machines in the manufacturing system of printed circuit boards. The inspection procedure for the solder joint defect type classification consists of an offline stage, which sets the optimal feature extraction region, and an online stage which classifies a defect type. In the offline stage, we use an optimization technique, namely, the genetic algorithm, to optimize the feature extraction region. In this stage, the optimal feature extraction region for defect type classification is constructed automatically. In the online stage, feature extraction regions are used to segment the solder joint image after component image acquirement. We then extract various color features from the segmented feature extraction regions. Next, we use support vector machine, which is one of the machine learning model’s method to classify the solder joint defect type. To evaluate the performance of the proposed method, ten types of solder joint defects were used in an experiment. The experimental results verified the effectiveness of the method in terms of the recognition rate, and its convenience.

Accuracy of cone-beam computed tomography, dental magnetic resonance imaging, and intraoral radiography for detecting peri-implant bone defects at single zirconia implants-An in vitro study.

To evaluate the diagnostic value of cone-beam computed tomography (CBCT), intraoral radiography (IR), and dental magnetic resonance imaging (dMRI) for detecting and classifying peri-implant bone defects at zirconia implants. Forty-eight zirconia implants were inserted in bovine ribs, 24 of which had standardized defects (1-wall, 2-wall, 3-wall, 4-wall) in two sizes (1 and 3mm). CBCT, IR, and dMRI were performed and analyzed twice by four readers unaware of the nature of the defects. Cohen's and Fleiss' kappa (κ), sensitivity, and specificity were calculated for the presence/absence of bone defects, defect size, and defect type. Cochran's Q-test with post hoc McNemar was used to test for statistical differences. A high intra- and inter-reader reliability (κ range: 0.832-1) and sensitivity/specificity (IR: 0.97/0.96; CBCT: 0.99/1; dMRI: 1/0.99) for bone defect detection were observed for all three imaging methods. For defect type classification, intra- (κ range: 0.505-0.778) and inter-reader (κ: 0.411) reliability of IR were lower compared to CBCT (κ range intrareader: 0.667-0.889; κ inter-reader: 0.629) and dMRI (κ range intrareader: 0.61-0.832; κ inter-reader: 0.712). The sensitivity for correct defect type classification was not significantly different for CBCT (0.81) and dMRI (0.83; p=1), but was significantly lower for IR (0.68; vs. CBCT p=0.003; vs. dMRI p=0.004). The sensitivity advantage of CBCT and dMRI for defect classification was smaller for 1-mm defects (CBCT/dMRI/IR: 0.68/0.72/0.63, no significant difference) than for 3-mm defects (CBCT/dMRI/IR: 0.95/0.94/0.74; CBCT vs. IR p=0.0001; dMRI vs. IR p=0.003). Within the limitations of an in vitro study, IR can be recommended as the initial imaging method for evaluating peri-implant bone defects at zirconia implants. CBCT provides higher diagnostic accuracy of defect classification at the expense of higher cost and radiation dose. Dental MRI may be a promising imaging method for evaluating peri-implant bone defects at zirconia implants in the future.

An inspection and classification method for chip solder joints using color grads and Boolean rules

In order to improve the comprehensive performance of solder joints inspection in three aspects, i.e. high recognition rate, detailed classification of defect types and fast inspection speed, a new detection and classification algorithm of the chip solder joints based on color grads and Boolean rules is developed in this paper. Firstly, the region features, evaluation features and color grads’ features are defined and extracted based on the special solder joint image, which is acquired by a particular image acquisition system composed of a 3-CCD color digital camera and a 3-color (red, green, and blue) hemispherical LED array illumination. Secondly, the models of solder joint types are built based on extracted features and statistical characteristics of solder joint types. Thirdly, the detection and classification method is designed and presented using Boolean rules, then eight common solder joint types, including the acceptable solder joint, pseudo, no solder, lacked solder, excess solder, shifted, tombstone, and miss component, can be classified and detected by the proposed algorithm. Fourthly, the proposed algorithm is optimized to improve the inspection speed based on a parallel computing method. Finally, to evaluate the performance of the proposed method, 79 pieces of PCBs with defects were inspected by the commercial AOI system developed by the authors which integrates the proposed algorithm. Experiment and result analysis illustrates that the proposed method is better than other methods in three aspects, it can detect and classify properly all the eight common types of solder joints, its detailed classification, and high correct rate, which is up to 97.7%, are more useful to the quality control in the manufacturing process, and its inspection speed is faster, thus helping us to improve the efficiency of the manufacturing process.

Weld Defect Classification in Radiographic Film Using Statistical Texture and Support Vector Machine

Weld defect identification requires radiographic operator experience, so the interpretation of weld defect type could potentially bring subjectivity and human error factor. This paper proposes Statistical Texture and Support Vector Machine method for weld defect type classification in radiographic film. Digital image processing technique applied in this paper implements noise reduction using median filter, contrast stretching, and image sharpening using Laplacian filter. Statistical method feature extraction based on image histogram was proposed for describing weld defects texture characteristic of a radiographic film digital image. Multiclass Support Vector Machine (SVM) algorithm was used to perform classification of weld defects type. The result of classification testing shows that the proposed method can classify 83.3% correctly from 60 testing data of weld defects radiographic films.

Comparative Analysis of the Mollon-reffin Minimal Colour Vision Test in Visually ‘Normal’ and Acquired Ocular Disease

Aim: Currently a variety of different colour vision tests exist in clinical practice each with its own merits and limitations. The Ishihara test is commonly used to test for congenital deficiency, causing red -green deficits. However acquired deficiency is also common in any ocular pathology affecting the optic nerve. The aim of this study is to compare the Mollon-reffinMinimal Test (MRMT) against the Ishihara and City colour vision tests. Methods: All three t ests were performed within the same clinic visit with a small rest between tests. The main parameters of interest were correlation levels between tests for defect detection, severity grading and defect type classification. We assessed the time taken to complete testing and evaluated the patients experience of test.

An integrated signal processing and neural networks system for steam generator tubing diagnostics using eddy current inspection

The primary purpose of this research was to develop an integrated approach by combining information compression methods and artificial neural networks for the monitoring of plant components using nondestructive evaluation (NDE) data. Specifically, data from eddy current inspection of steam generator tubing were utilized to evaluate this technology. The focus of the research was to develop and test various data compression methods (for eddy current data) and the performance of different neural network paradigms for defect classification and defect parameter estimation. Feedforward, fully-connected neural networks, that use the back-propagation algorithm for network training, were implemented for defect classification and defect parameter estimation using a modular network architecture. A large eddy current tube inspection database was acquired from the Metals and Ceramics Division of Oak Ridge National Laboratory (ORNL). These data were used to study the performance of artificial neural networks for defect type classification and for estimating defect parameters. Most of the study was made using the NeuralWorks Professional II/Plus software. A PC-based data pre-processing and display program was also developed as part of an expert system for data management and decision making. The results of the analysis showed that for effective (low-error) defect classification and estimation of parameters, it is necessary to identify proper feature vectors using different data representation methods. The integration of data compression and artificial neural networks for information processing was established as an efficient technique for automation of diagnostics using nondestructive evaluation methods.