Detection Of Multiple Defects Research Articles

Ultrasonic guided wave defect detection method for tank bottom plate based on SH0 mode multichannel magnetostrictive sensor

To address the challenge of ultrasonic detection of multiple defects in large-scale tank bottom plates, a multi-channel magnetostrictive sensor was designed and developed to effectively excite SH (Shear Horizontal) mode guided waves. The optimal overall scheme of the sensor was obtained by considering the sensor excitation energy, coupling efficiency, and fabrication process. Through experimental tests, the consistency of each channel of the sensor is evaluated, and the center frequency and detection range of the sensor are measured. To detect defects of varying sizes and depths in steel plates, a rotating total focus imaging method is proposed. And we conducted experiments using a multi-channel magnetostrictive sensor to achieve SH0 guided wave imaging of multiple defects across the full domain. Our experimental results demonstrate that, across 8 rotations, we achieved a minimum relative position error of 4.32% and a minimum absolute angle error of 2.16°. Our research provides an effective solution for ultrasonic detection of tank bottom plates, and the proposed method is also applicable for defect detection in other large plate-like structures.

An automated defect detection method for optimizing industrial quality inspection

Detecting defects in industrial-quality inspection is an important task, but defect detection remains challenging due to limited size, various types of defects, and imbalanced samples in images. In this paper, we propose a novel automatic defect detection network that efficiently detects defects. We introduce the Automated Sample Assignment (SCID) algorithm, which automatically divides the distribution of positive and negative samples based on similarity scores using the Gaussian mixture model. The Automated Sample Assignment (SCID) algorithm solves the problem of sample imbalance in defect detection networks caused by manually setting the intersection over the union (IOU) threshold and assigning positive and negative labels to samples. Then, instead of the pooling method used by most backbone networks, we use a loss fusion pool named LE-Pool to retain more detailed information. Finally, we use the soft weighted attention mechanism of the detection head (S-head) to identify noise information. Experimental results on multiple defect detection datasets show that the proposed method is superior to other detection methods in terms of accuracy.

Multiple Defect Classification Method for Green Plum Surfaces Based on Vision Transformer

Green plums have produced significant economic benefits because of their nutritional and medicinal value. However, green plums are affected by factors such as plant diseases and insect pests during their growth, picking, transportation, and storage, which seriously affect the quality of green plums and their products, reducing their economic and nutritional value. At present, in the detection of green plum defects, some researchers have applied deep learning to identify their surface defects. However, the recognition rate is not high, the types of defects identified are singular, and the classification of green plum defects is not detailed enough. In the actual production process, green plums often have more than one defect, and the existing detection methods ignore minor defects. Therefore, this study used the vision transformer network model to identify all defects on the surfaces of green plums. The dataset was classified into multiple defects based on the four types of defects in green plums (scars, flaws, rain spots, and rot) and one type of feature (stem). After the permutation and combination of these defects, a total of 18 categories were obtained after the screening, combined with the actual situation. Based on the VIT model, a fine-grained defect detection link was added to the network for the analysis layer of the major defect hazard level and the detection of secondary defects. The improved network model has an average recognition accuracy rate of 96.21% for multiple defect detection of green plums, which is better than that of the VGG16 network, the Desnet121 network, the Resnet18 network, and the WideResNet50 network.

The Steel Surface Multiple Defect Detection and Size Measurement System Based on Improved YOLOv5

In the process of steel production, the defects on the surface of steel will adversely affect the subsequent processing of a product. Accurate detection of such defects is the key to improve production efficiency and economic benefits. In this paper, an end-to-end steel surface defect detection and size measurement system based on the YOLOv5 model is designed. Firstly, in consideration of the defect location and direction correlation in the production process, a coordinate attention mechanism is added at the head of YOLOv5 to strengthen the spatial correlation of the steel surface and an adaptive anchor box generation method based on defect shape difference feature is proposed, which realizes the detection of three main types of defects on the Pytorch deep learning framework. Secondly, BiFPN is used to strengthen the feature fusion and a transformer encoder is added to improve the performance of detecting small defects. Thirdly, calculate the conversion ratio between the pixel and the actual size according to the standard reference specimen and obtain the actual size through the pixel statistics of the defect area to achieve pixel level size measurement. Finally, the steel surface defect detection and size measurement system are designed in this paper, which consist of various hardware, related measurement, and detection algorithms. According to the experimental results, the comprehensive defect detection accuracy of this method reaches 93.6%, of which the scratch detection accuracy reaches 95.7%. The detection speed reaches 133 fps and the defect size measurement accuracy reaches 0.5 mm. Experimental result shows that the defect detection and size measurement system designed in this paper can accurately detect and measure various industrial production defects and can be applied to the actual production process.

Robust Anomaly Map Assisted Multiple Defect Detection with Supervised Classification Techniques

Industry 4.0 aims to optimize the manufacturing environment by leveraging new technological advances, such as new sensing capabilities and artificial intelligence. The Discriminatively trained Reconstruction Anomaly Embedding Model (DRAEM) technique has shown state-of-the-art performance for unsupervised classification. The ability to create anomaly maps (images highlighting areas where defects probably lie) can be leveraged to provide cues to supervised classification models and enhance their performance. Our research shows that the best performance is achieved when training a defect detection model by providing an image and the corresponding anomaly map as input. Furthermore, such a setting offers consistent performance when framing defect detection as a binary or multiclass classification problem and is unaffected by class balancing policies. We performed the experiments on three datasets with real-world data provided by Philips Consumer Lifestyle BV.

Surface Defect Detection Method Based on Improved Semisupervised Multitask Generative Adversarial Network

The detection methods based on deep learning networks have attracted widespread interest in industrial manufacture. However, the existing methods are mainly trapped by a large amount of training data with excellent labels and also show difficulty for the simultaneous detection of multiple defects in practical detection. Therefore, in this article, a defect detection method based on improved semisupervised multitask generative adversarial network (iSSMT-GAN) is proposed for generating better image features and improving classification accuracy. First, the training data are manually labeled according to the types of defects, and the generative adversarial network (GAN) is constructed according to the reliable annotations about defects. Thus, a classification decision surface for the detection of multitype defects is formed in the discriminative network of GAN in an integrated manner. Moreover, the semisupervised samples generated by the discriminative network give the generative network feedback for enhancing the image features and avoiding gradient disappearance or overfitting. Finally, the experimental results show that the proposed method can generate high-quality image features compared with the classic GAN. Furthermore, this increase in classification accuracy of RegNet model, MobileNet v3 model, VGG-19 model, and AlexNet-based transfer learning is 3.13%, 2.30%, 2.48%, and 3.12%, respectively.

AE-RTISNet: Aeronautics Engine Radiographic Testing Inspection System Net with an Improved Fast Region-Based Convolutional Neural Network Framework

To ensure safety in aircraft flying, we aimed to use deep learning methods of nondestructive examination with multiple defect detection paradigms for X-ray image detection. The use of the fast region-based convolutional neural network (Fast R-CNN)-driven model was to augment and improve the existing automated non-destructive testing (NDT) diagnosis. Within the context of X-ray screening, limited numbers and insufficient types of X-ray aeronautics engine defect data samples can, thus, pose another problem in the performance accuracy of training models tackling multiple detections. To overcome this issue, we employed a deep learning paradigm of transfer learning tackling both single and multiple detection. Overall, the achieved results obtained more than 90% accuracy based on the aeronautics engine radiographic testing inspection system net (AE-RTISNet) retrained with eight types of defect detection. Caffe structure software was used to perform network tracking detection over multiple Fast R-CNNs. We determined that the AE-RTISNet provided the best results compared with the more traditional multiple Fast R-CNN approaches, which were simple to translate to C++ code and installed in the Jetson™ TX2 embedded computer. With the use of the lightning memory-mapped database (LMDB) format, all input images were 640 × 480 pixels. The results achieved a 0.9 mean average precision (mAP) on eight types of material defect classifier problems and required approximately 100 microseconds.

Three-dimensional images generated from diffuse ultrasound wave: detections of multiple cracks in concrete structures

The simultaneous detection of multiple defects in concrete structures is a task of pivotal importance for non-destructive testing and evaluation. Diffuse waves experiencing multiple scattering inside media are demonstrated to be sensitive to weak defects. Here, an analytic model is presented for diffuse wave decorrelation associated with sensitivity kernel that describes the time-of-flight distribution in strongly scattering environments. The model is then used for generating three-dimensional images that involve estimating perturbations at each localized position through an iterative, non-linear algorithm. With the consideration of loads and micro-cracks effects on diffuse waves, an application of the approach to a real-size concrete beam shows features that denote the positions and depths of multiple existing cracks. Extension of the approach to other strongly scattering media such as tissues and volcanos is straightforward. This study offers great potential for practical applications such as structural health monitoring, medical image generation, and seismic monitoring.

A Deep-Learning-Based Multiple Defect Detection Method for Tunnel Lining Damages

Diagnosing tunnel lining structural damages is vital to ensure safe tunnel operations. However, the detection of multiple defect is challenging task due to the size imbalance between cracks, spalling, and backgrounds. Currently, deep-learning-based methods for multiple defect are dependent on multiple-stage networks, which have limited their scalability and complex frame working processes. To accurately recognize the multiple defect at the pixel-level using only one-stage networks, a new method was proposed, which integrated the basic SegNet with a focal loss function, and was referred to as an FL-SegNet method. The focal loss function was adopted to address the problem of the size imbalance by down-weighing the losses assigned to the well-classified samples, and then the training was focused on the hard samples. Furthermore, comparative experiments were performed to evaluate the performances of the different methods. The experimental results demonstrated that FL-SegNet method was capable of accurately predicting the profiles of small-sized cracks and overlapping damages even under various noise conditions, and successfully outperformed the two-stream method and the basic SegNet method in this regard. The performance metrics (MPA and MIoU) of the FL-SegNet method were significantly higher than those of other multiple defect detection approaches in different scenarios (images with small-sized damages attained to 81.53% and 69.86%, increased by 11.99 % and 4.88% compared with two-stream method, and increased by 17.78% and 7.69% compared with basic SegNet). Therefore, this paper provides an effective solution for the future detection of multiple defect in tunnel linings.

Incidence and natural history of neonatal isolated ventricular septal defects: Do we know everything? A 6-year single-center Italian experience follow-up.

Despite ventricular septal defects (VSDs) are the most common congenital heart diseases (CHDs) in the neonatal period, their incidence and natural history are still debated and their follow-up and management strategies remain controversial. Our aim was to evaluate the incidence and natural history of isolated VSDs. From January 1996 to December 2015 all neonates with a CHD suspicion were referred to the Cardiological Department of Grosseto Misericordia Hospital. Only newborns with confirmed isolated VSD were enrolled in this study and followed for 6 years. Our 343 newborns with an isolated VSD (incidence of 10.45/1000/births) account for 64% of all detected CHDs. VSDs location were as follows: muscular (73.8%), perimembranous (11.3%), inlet (1%), and outlet (0.8%). Of the located VSDs, 90% were small, 7.5% moderate, and 2.5% large, respectively. Spontaneous closure was observed in 96 (29.2%) of the VSD patients at 6-month, 198 (60.2%) at 1-year, 261 (79.3%) at 2-year, and in 302 (91.8%) at 6-year follow-up. Risk factors for defect persistence were a perimembranous location (P = .001; HR: 0.508, CI: 0.342-0.755), detection of multiple defects (P = .043; HR: 0.728, CI: 0.536-0.990), and male gender (P < .048; HR: 0.783, CI: 0.615-0.998), respectively. We here provide an incidence and natural history of neonatal isolated VSDs in a neonatal Caucasian population. These data may be useful for the development of expert consensus/standard recommendation guidelines for the follow-up and VSD management, data that are currently lacking.

Local interrogations for detection of multiple defects using an active sensor network

With the application of a clock-like sensor network, a local interrogation strategy is established to extend the reconstruction algorithm for the probabilistic inspection of damage (RAPID) into multiple defects identification. In this strategy, the monitoring area is divided into several monitoring subareas, and each subarea is covered by a particular sensor network. Local interrogations are then employed in each subarea using the associated network of sensing paths. Benefit from that, the artifacts arise from the intersection of key sensing paths associated with different defects may be absent completely from some monitoring subarea. With the application of a proper image fusion strategy, the artifacts would be reduced and the damage indications could be preserved. The algorithm is applied to an aluminum plate with two artificial through-thickness holes introduced. The results demonstrated that the RAPID algorithm with the application of local interrogations is capable of identifying multiple defects in plate-like structures.

Analysis on Characteristic Features of L-Mode Guided Wave in Tube Testing

Applications of long range ultrasonic testing are on cylindrical products(pipes and tubes) and planar products(plates and sheets).Guided wave testing using circumferential and longitudinal modes are getting more and more importance as there are huge requirements for testing pipes and tubes in power generators, Chemical plants, steam generators etc.. There will be infinite number of modes with charectestic propagation and wave structure. This propagation characteristic and wave structure plays a major role in deciding its suitability for use in specific application. It is highly essential to use mode control and to understand the generated modes and its characteristics based on which only testing adequacy can be evaluated. In this paper the characteristic features of L-mode generated using a source design parameter for tube testing are explained. The dispersion characteristic of L-mode are discussed using constructed Dispersion Curves, phase and group velocity, its multiple defect detection capability, pulse shape and amplitude dynamics with propagation range and inclination. The details of controlled experiment conducted and the observations are discussed. The derived inferences are explained based on the applicable particle displacement equations and dispersion characteristic equations.

Integral-FEM Simulation and Independent Component Analysis for Multiple Defect Separation from Pulse Eddy Currents Signals

Pulse eddy currents simulated using an integral formulation and Independent Component Analysis (ICA) for signal separation are used for detection of multiple defects in thick conductive plates. Numerical results for the eddy-currents signals, and signal separation using ICA, are shown.

The Grand Pareto: A Methodology for Identifying and Quantifying Yield Detractors in Volume Semiconductor Manufacturing

A method of communicating a unified pareto, we call "Grand Pareto," for technology-wide failure mechanisms that limit the profitability of a fabrication facility is presented. The Grand Pareto leverages multiple defect detection and isolation techniques in conjunction with state-of-the-art physical failure analysis and statistical yield analysis techniques to create a single message for the process community to drive the yield improvement efforts. The methodology has been successfully deployed at IBM where it has been assisting identification of key yield detractors for several high-end microprocessors in volume production

Multicrack Detection in Two-Dimensional Structures by Means of GA-Based Strategies

We propose a methodological approach for the detection of multiple defects inside dielectric or conductive media. Two innovative algorithms are developed starting from the inverse scattering equations solved by means of different optimization strategies. In the first implementation, a hierarchical strategy based on parallel-subprocesses is considered, whereas the second algorithm employs a single-process architecture. Whatever the implementation, the arising cost function is minimized through a suitable hybrid-coded genetic algorithm, whose individuals encode the problem unknowns. In order to achieve a computational saving, the formulation based on the inhomogeneous Green's function is adopted and each crack-region is parametrized by means of a selected set of descriptive parameters. The approach as well as its different implementations are assessed through a selected set of numerical experiments and in comparison with previously developed single-crack inverse scattering methods

Nondestructive Testing Method for Finding out the Defects in a Composite Liner

A composite liner of carbon phenolic has been inspected by ultrasonic, X-ray radiography andX-ray computed tomography (CT) to find defects like delaminations, debonds, voids, foreign inclusions, etc.The geometry, detection of multiple defects and porosity of the liner make ultrasonic testing (pulse-echo anddrycoupling) difficult for inspection. X-ray radiography being a non-contact technique finds multiple defectsbut compresses the structural information of 3-D volume into a 2-D image and interferes with overlyingand underlying areas of the object. X-ray CT generates an image of a thin and cross-sectional slice ofan object. R e linear attenuation coefficients in terms of Hounsfield values have been measured, comparedand correlated with CT images at the contrasts observed. 3-D images can be generated by stacking2-D cross-sectional images of the slices. These 3-D images can be cut at any angle of choice for mappingthe extent of delaminated/debonded areas. This type of information is difficult to obtain with conventionalnon-destructive testing techniques.

An integrated robotic and machine vision system for surface flaw detection and classification

Detection of multiple defects on a surface of specified material is an important application of machine vision in the manufacturing environment. Several procedures have been developed to detect the existence of defects such as scratches, cracks or bubbles which may occur during production processes and handling of materials such as glass and plastics. However, few procedures classify the determined defect. This paper presents an integrated robotic and machine vision system along with an algorithm for detection/classification of both multiple defects and the color quality of the inspected material.

Automated visual inspection : Batchelor, B.G.; Hill, D.A.; Hodgson, D.C. (eds.) IFS (Publications) Ltd., 561 pp. (1985)

Existing methods of inspection are often inadequate and are the main limitation to improvement. Automated visual inspection represents a significant contribution to the second industrial revolution. This book discusses the automation of inspection with light and its various industrial applications. Automated visual inspection is taken to encompass a wide variety of tasks, including blemish detection, complex patterned-surface examination, single and multiple defect detection, measurement of tools and components, counting, grading, sorting, sizing, orientation and posture recognition (for robot control), in process monitoring and control. Visual inspection can also be extended to cover other situations where images are generated by thermography, radiography, fluoroscopy, NMR, and ultrasonics. These are discussed, as are some of the non-imaging uses of light.