Defect Inspection Research Articles

Megahertz detection of spectroscopic polarization by a time-encoded supercontinuum vector beam.

We demonstrated a 40-MHz detection of spectroscopic polarization by a supercontinuum vector beam with a wavelength-dependent polarization state. To achieve the high-repetition-rate measurement, we detected the rotation angle of polarization and the spectrum by measuring the temporal waveform using a photodetector after expanding the pulse duration of the supercontinuum vector beam. The spectrum of the supercontinuum vector beam was measured using a spectrometer. We compared it with the temporal waveforms, confirming a good agreement of spectra between the conventional spectrometer and the temporal waveforms. The detection method is useful for many applications requiring high-repetition-rate spectroscopic-polarization measurements, such as the defect inspection of thin optical materials.

Rapid damage reconstruction imaging of composite plates using non-contact air-coupled Lamb waves

Air-coupled ultrasonic Lamb wave damage imaging is a method for rapid, non-contact damage detection of plate specimens. However, the current technique based on Lamb waves has limitations regarding accuracy and quality of imaging. To address this issue, an improved bilateral boundary-enhanced reconstruction algorithm for the probabilistic inspection of defects (BBERAPID) is proposed. This algorithm optimizes the probability distribution near the damage boundary to enhance the accuracy of damage imaging. The skewed bilateral boundary distribution algorithm can improve the probability distribution of edges on both sides of the damage and alleviate the problem of abnormal boundary expansion. In addition, a new damage index (DI) calculation method is proposed to accelerate the DI trend near the damage. The new DI can improve the fineness of the damage boundary, thus better distinguish the healthy area from the damage area. Traditional DI has the problem of abnormal distribution at the damage boundary, which BBERAPID solves. The probability distribution near the damage is optimized to improve the quality and accuracy of damage imaging. Experimental studies are performed using composite plates to validate the proposed method further. The effectiveness and accuracy of the proposed BBERAPID algorithm for detecting damages in composite plates are verified by comparing it with existing imaging methods through experimental validation. The proposed non-contact air-coupled Lamb wave damage imaging method is suitable for rapidly and accurately characterizing defects in composite plates.

Enhanced YOLOv8 with BiFPN-SimAM for Precise Defect Detection in Miniature Capacitors

In the domain of automatic visual inspection for miniature capacitor quality control, the task of accurately detecting defects presents a formidable challenge. This challenge stems primarily from the small size and limited sample availability of defective micro-capacitors, which leads to issues such as reduced detection accuracy and increased false-negative rates in existing inspection methods. To address these challenges, this paper proposes an innovative approach employing an enhanced ‘you only look once’ version 8 (YOLOv8) architecture specifically tailored for the intricate task of micro-capacitor defect inspection. The merging of the bidirectional feature pyramid network (BiFPN) architecture and the simplified attention module (SimAM), which greatly improves the model’s capacity to recognize fine features and feature representation, is at the heart of this methodology. Furthermore, the model’s capacity for generalization was significantly improved by the addition of the weighted intersection over union (WISE-IOU) loss function. A micro-capacitor surface defect (MCSD) dataset comprising 1358 images representing four distinct types of micro-capacitor defects was constructed. The experimental results showed that our approach achieved 95.8% effectiveness in the mean average precision (mAP) at a threshold of 0.5. This indicates a notable 9.5% enhancement over the original YOLOv8 architecture and underscores the effectiveness of our approach in the automatic visual inspection of miniature capacitors.

Deep Fuzzy Cognitive Maps for Defect Inspection in Antenna Assembly

Regardless of the tremendous technological development in manufacturing processes and equipment, assembly errors are an emerging concern. Hence, quality inspection software is of high importance for capturing defects at the source and preventing further error propagation. The advantages of Fuzzy Cognitive Maps (FCMs) as knowledge representation models with explainable capabilities can be compounded with the advantages of Deep Convolutional Neural Networks (CNNs) in order to acquire accurate and interpretable inspection results in Smart Manufacturing. In this work, an explainable binary classification approach is performed, namely DeepK-FCM, and appears to be able to assess sufficiently industrial images from TELEVES. The suggested DeepK-FCM methodology, inheriting the powerful characteristics of deep learning and FCMs, includes a number of steps as follows: fine-tuning of well-known CNNs for feature extraction with transfer learning, feature clustering by performing K-Means algorithm, definition of causal relationships through fuzzy measures on similarities produced and FCM model for the decision-making. Through the experimental analysis on a real industrial data set, it is being proved that the current approach with DeepK-FCM is more efficient for the task of defect inspection in the antenna assembly than the straightforward binary classification accomplished by CNNs. The attained accuracy of 80% is improved by 3% compared to the state-of-art CNN classifiers and indicates a high potential even when the data are scarce. In addition, the decision-making with FCMs levitates the interpretability of the system which contributes to the efforts towards a more explainable AI method for quality control inspection.

AI-driven real-time failure detection in additive manufacturing

The optimisation of 3D printing parameters for manufacturing biomedical devices is an emerging interdisciplinary field that incorporates artificial intelligence techniques such as machine learning and deep learning. In this particular study, the focus is on the fabrication of biocompatible finger splints using digital light processing 3D printing technology, followed by UV curing, to evaluate their quality. By leveraging vibration data from printers, which cannot be captured through visual inspection of layer defects, this study aims to develop a predictive model for assessing the failures of printed parts. Here, a closed-loop detection system is proposed to identify failure phenomena in 3D resin printing, combining both cloud and edge computing technologies to effectively detect and address potential failures in the printing process.

Defect Detection and Closed-loop Feedback Using Machine Learning for Fused Filament Fabrication

The objective of this study was to develop a closed-loop system for a commercial fused filament fabrication printer based on visual machine learning inspection of common defects. Convolutional neural network was used to identify levels of common defects: stringing, over/under-extrusion, and weak infill. Transfer learning was used to adapt a pre-trained model to fit this problem, as it involves incrementally fine-tuning the model parameters to new data. The observation model achieved an accuracy of 92.86% on validation data set and 90.0% on the testing data set. By modifying the input G-code, the custom program could adjust the feed-rate, nozzle temperature, material extrusion amount, and fan speed to correct for identified extrusion defects.

A Visual Inspection and Classification Method for Camshaft Surface Defects Based on Defect Similarity Measurement

A Visual Inspection and Classification Method for Camshaft Surface Defects Based on Defect Similarity Measurement

Automatic Optical Inspection of Microscale Defects on Curved Surface

Automatic Optical Inspection of Microscale Defects on Curved Surface

Evaluation of different surface topography measurement methodologies for characterizing weld beads in shipbuilding

The weld bead is crucial in identifying defects during laser welding, aiding in the evaluation of weld strength and quality. However, accurately monitoring and identifying defects with cameras can be challenging in the preassembly phase. This paper compares surface topography of laser weld beads using in-process Optical Coherence Tomography (OCT) and post-process Focus Variation Microscopy (FVM). Both techniques were employed on welded A36 structural steel plates to detect the surface defects. The weld profile was evaluated according to EN ISO 13919-1, and the quality level B was given to the surface topography measurement from both techniques. The correspondence between these two methodologies is imperative for implementing profile data in advanced monitoring systems. Thus, by reducing the delay between defect formation and inspection, this paper supports a more intelligent welding process, particularly in the shipbuilding industry.

Simulation study of decoherence and light intensity uniformization for extreme ultraviolet of uniform light pipe

<sec>Free electron laser (FEL) is a high-quality laser source with wavelengths ranging from short-wave X-ray to long-wave infrared ray. Extreme ultra-violet (EUV) radiation at <i>λ</i> = 13.5 nm emitted by FEL can be used in manufacturing integrated circuit, such as EUV lithography exposure and mask defect inspection. However, the high spatial coherence characteristics and similar Gauss intensity profile distribution of FEL source has a negative effect on imaging, and cannot meet the requirements of imaging applications in EUV lithography. In this work, a newly light pipe for decoherence and intensity uniformity in an EUV spectral range is designed through the simulation calculations. The new light pipe consists of two pairs of tilted elements which are symmetrically distributed in the <i>y</i>-<i>z</i> plane and <i>x</i>-<i>z</i> plane, respectively. In this way, the beam transmission divergence in two dimensions can be widened at the same time, and the disturbance of the ray transmission track and spatial phase distribution is increased, so as to achieve the uniformization of light intensity and the reduction of spatial coherence.</sec><sec>The simulation results show that for an EUV Gaussian beam at <i>λ</i> = 13.5nm, with a diameter of 200 μm, and a divergence of 20 mrad, the newly designed light pipe has more significant decoherence and illumination field uniformity than the conventional light pipe structure. When the new light pipe has an inner diameter of 1mm, a total length of not less than 600 mm, and a tilt angle of 10mrad , a basically uniform illumination field can be obtained, the coherence is completely disordered, and the non-uniformity of light intensity distribution in the illumination field decreases to 0.97 from 28.38 achieved by conventional cylindrical light pipe. At the same time, the light power transmission efficiency is about 37.6% and the maximum transmission efficiency is about 44.58%. The uniformity can be further improved by increasing the number of reflections. When the inner diameter and tilted angle of the light pipe are unchanged, the length of the light pipe increases to 1m, the non-uniformity of intensity distribution at the illumination field further decreases to 0.90, and the light power transmission efficiency is about 22.35%. The results show that the newly designed light pipe structure can meet the application requirements of decoherence and improve the uniformity of illumination field at EUV wavelength range, and it has great application prospects in EUV lithography and other imaging applications.</sec>

A Lightweight Fine-Grained Perception Defect Inspection Network Under Multitask Learning on Highly Reflective Surface

A Lightweight Fine-Grained Perception Defect Inspection Network Under Multitask Learning on Highly Reflective Surface

Performances of an in-line deep learning-based inspection system for surface defects of die-cast components for hybrid vehicles

The detection of surface defects is a critical activity for quality assurance of automotive products. Deep learning (DL) methods range among the most suitable tools to support this task. They allow the inspection system to learn how to detect surface defects on the basis of a series of sample images. However, most of the past studies regard laboratory experiments. This study demonstrates the actual integration of an inspection system into an industrial production plant to perform surface defect inspection operations on machined and unmachined areas of aluminium components for hybrid cars. The proposed DL method is practical for industry, where the number of available defect samples is low. Experimental results showed a True Positive Rate of 98%.

Characterization of Multimodal Spot Scanning Imaging System for Wafer Defect Inspection

Characterization of Multimodal Spot Scanning Imaging System for Wafer Defect Inspection

Enhancing Battery Exterior Defect Inspection Accuracy Through Defect-Background Separated GAN Development

Enhancing Battery Exterior Defect Inspection Accuracy Through Defect-Background Separated GAN Development

Cross-factory Polarizer Sheet Surface Defect Inspection System Based on Multi-Teacher Knowledge Amalgamation

Cross-factory Polarizer Sheet Surface Defect Inspection System Based on Multi-Teacher Knowledge Amalgamation

Enhancing RCF rail defect inspection on the Serbian railway network

The interaction between the wheel and the running rails within the railway system introduces intricate stress patterns, resulting in the formation of rolling contact fatigue (RCF) rail defects. The magnitude of this stress is contingent upon factors such as track performance, vehicle characteristics, and service conditions. While advancements in rail metallurgy can mitigate the issue to some extent, no economically viable steel composition currently exists that can completely withstand the repetitive stresses associated with RCF. It is more cost-effective to properly maintain rails for longer use rather than replace them entirely. The paper emphasizes the importance of classifying and coding RCF rail defects in light of their potential adverse effects on rail transport safety. It provides an analysis of the available inspection methods for RCF rail defects and recommends the ones that should be implemented on the Serbian railway network. A combination of proposed inspection methods is preferred to increase detection efficiency for different types of RCF defects.

Improving the Navigation System of Underground Pipe Defect Detection Device

This article addresses the issue of noise and drift in microelectromechanical gyroscopes and their impact on measurement accuracy in engineering applications. The use of a complementary filter is proposed to combine information from the accelerometer and gyroscope to reduce inaccuracies. Research shows that the accelerometer has better result repeatability, which is important for obtaining stable measurements. At the same time, the gyroscope may be more effective in measuring translational movements. The selection of sensor sensitivities and proper parameter tuning are crucial aspects. A developed system is capable of effectively filtering and measuring the angle of an object, and the use of a complementary filter improves measurement accuracy. The proposed approach can be successfully utilized for accurately detecting the angle of position of a measurement setup in defect inspection of underground pipelines.

Learning-based defect inspection of laptop covers with a laser profile sensor

Quality inspection is a crucial process in the manufacturing industry, which significantly impacts the productivity and profits of a factory. To reduce costs and enhance quality control, companies are increasingly driving automation of this process due to the limitations of manual labour. However, the lack of flexible solutions that can adapt to the changing environment in factories has been a challenge. Computer-vision-based inspection methods often require hefty mechanisms and strict lighting conditions, which limit the investment in such solutions. Furthermore, with increasing demands for precision, traditional algorithms are struggling to cope with the problem, while data-driven solutions like deep learning require significant effort in acquiring and labelling data. To address these problems, this paper proposes an eye-in-hand system, i.e., a sensor attached to a robotic manipulator, integrated with unsupervised learning algorithms. One typical product, laptop bottom covers, is tested and discussed with regard to the system to validate our method since it has tiny defects and light-absorbing coatings. With only nine samples of the product, our system solved the case and demonstrated high-precision results and a flexible workflow. In this paper, our main contributions are: 1) we propose a flexible quality inspection system and workflow that needs extremely small volumes of data with unsupervised learning algorithms; and 2) we first propose using a laser profile sensor for collecting 2D intensity images, that are resilient to ambient light disturbance, of matt laptop cover data without lighting equipment.

Adaptive explainable artificial intelligence for visual defect inspection.

Explainable Artificial Intelligence promises to deliver means so that humans better understand the rationale behind a particular machine learning model. In the image domain, such information is frequently conveyed through heat maps. Along the same line, information regarding defect detection for unsupervised methods applied to images can be conveyed through anomaly maps. Nevertheless, heat maps or anomaly maps can convey inaccurate information (artifacts), or their perceptions may differ across different persons. Therefore, the user experience could be enhanced by collecting human feedback and creating predictive models on how these could be recolored to bridge the gap between the original heat maps and anomaly maps created with explainability techniques and the output expected by humans. We envision this work as relevant in at least two scenarios. First, enhance anomaly and heat maps when conveying information regarding machine vision models deployed in production to remove information deemed unnecessary by the user but systematically present through the explainability technique due to underlying model issues (artifacts). Second, adapt anomaly and heat maps based on users’ perceptual needs and preferences.

DES-YOLO: a novel model for real-time detection of casting surface defects.

Surface defect inspection methods have proven effective in addressing casting quality control tasks. However, traditional inspection methods often struggle to achieve high-precision detection of surface defects in castings with similar characteristics and minor scales. The study introduces DES-YOLO, a novel real-time method for detecting castings' surface defects. In the DES-YOLO model, we incorporate the DSC-Darknet backbone network and global attention mechanism (GAM) module to enhance the identification of defect target features. These additions are essential for overcoming the challenge posed by the high similarity among defect characteristics, such as shrinkage holes and slag holes, which can result in decreased detection accuracy. An enhanced pyramid pooling module is also introduced to improve feature representation for small defective parts through multi-layer pooling. We integrate Slim-Neck and SIoU bounding box regression loss functions for real-time detection in actual production scenarios. These functions reduce memory overhead and enable real-time detection of surface defects in castings. Experimental findings demonstrate that the DES-YOLO model achieves a mean average precision (mAP) of 92.6% on the CSD-DET dataset and a single-image inference speed of 3.9 milliseconds. The proposed method proves capable of swiftly and accurately accomplishing real-time detection of surface defects in castings.