Defect Images Research Articles

Study on intelligent recognition of urban road subgrade defect based on deep learning.

China's operational highway subgrades exhibit a trend of diversifying types and an increasing number of defects, leading to more frequent urban road safety incidents. This paper starts from the non-destructive testing of urban road subgrade defects using geological radar, aiming to achieve intelligent identification of subgrade pathologies with geological radar. The GprMax forward simulation software is used to establish multi-layer composite structural models of the subgrade, studying the characteristics of geological radar images for different types of subgrade diseases. Based on the forward simulation images of geological radar for subgrade defects and field measurement data, a geological radar subgrade defect image database is established. The Faster R-CNN deep learning algorithm is applied to achieve target detection, recognition, and classification of subgrade defect images. By comparing the loss value, total number of identified regions, and recognition accuracy as metrics, the study compares four improved versions of the Faster R-CNN algorithm. The results indicate that the faster_rcnn_inception_v2 version is more suitable for the intelligent identification of non-destructive testing of urban road subgrade defects.

Lightweight defect detection algorithm of tunnel lining based on knowledge distillation

Due to the influence of construction quality, engineering geology and hydrological environment, defects such as dehollowing and insufficient compaction can occur in tunnels. Aiming at the problems of complex detection model, poor real-time performance and low accuracy of the current tunnel lining defect detection methods, the study proposes a lightweight defect detection algorithm of tunnel lining based on knowledge distillation. Firstly, a high-precision teacher model based on yolov5s was constructed by constructing a C3CSFM module that combines residual structure and attention mechanism, a MDFPN network structure with multi-scale feature fusion and a reweighted RWNMS re-screening mechanism. Secondly, in the distillation process, the feature and output dimension results are fused to improve the detection accuracy, and the mask feature relationship is learned in the space and channel dimension to improve the real-time detection. Tests on the tunnel lining radar defect image dataset showed that the number of parameters of the improved model was reduced from 16.03 MB to 3.20 MB, a reduction of 80%, and the average accuracy was improved from 83.4 to 86.5%, an increase of 3.1%. On the basis of maintaining the structure and detection performance of the model, the lightweight degree of the model is greatly improved, and the high-precision and real-time detection of tunnel lining defects is realized.

Research on the Forward Simulation and Intelligent Detection of Defects in Highways Using Ground-Penetrating Radar

The increasing variety and frequency of subgrade defects in operational highways have led to a rise in road safety incidents. This study employed ground-penetrating radar (GPR) detection and forward simulation to analyze the characteristic patterns of common subgrade defects, such as looseness, voids, and cavities. Through the integration of instantaneous feature information from different defect patterns with complex signal techniques, the boundary judgment of structural layers and anomalies in GPR images of various subgrade defects was improved. An intelligent recognition platform was established, and a radar image dataset was created and trained to evaluate the recognition performance of the You Only Look Once (YOLO) v3 and Single-Shot Multi-Box Detector (SSD) algorithms. Evaluation metrics such as precision, recall, F1-score, average precision (AP), and mean average precision (mAP) were used to assess the detection efficiency and accuracy for subgrade defect images. The results showed that YOLO v3 achieved an average detection accuracy of 76.69%, while the SSD achieved 75.07%. This study demonstrates that the reliability of the intelligent recognition and classification of highway subgrade defects can be enhanced by using GPR for non-destructive testing.

High-resolution weld defect detection with RSU-MLP and dynamic kernel supervision

The challenges posed by high-resolution X-ray images of weld defects—including fuzzy features, multi-scale variability, small targets, and sample imbalance—create significant obstacles for accurate defect localization. Traditional deep learning methods typically emphasize global image characteristics, often neglecting local and multi-scale information, which leads to the inaccurate detection of small defects. To address this issue, we propose a novel high-resolution weld defect detection method called RSU-MLP. This method combines the RSU-MLP network with multi-scale feature extraction technology to effectively capture both local and global image features. Additionally, we design a dynamic kernel adaptive segmentation head based on a hybrid extended convolutional attention mechanism, which adaptively adjusts the receptive field size of the convolutional kernel, thereby enhancing detection capabilities for defects of varying scales. Furthermore, an in-depth supervision mechanism is introduced to monitor and learn defects at different levels of the network, leading to improved detection performance. Experimental results on real high-resolution welding datasets demonstrate that the proposed method achieves promising results in welding defect detection. Compared to traditional approaches, this method provides more accurate detection and identification of various welding defects, achieving DICE and Jaccard scores of 75.97% and 64.01%, respectively. This represents the best performance, demonstrating both robustness and generalization capabilities. Consequently, it offers a reliable automated detection method for producing high-quality welded products.

Research on the veneer defect image enhancement algorithm based on AMEF-AGC.

Aiming at the veneer defect image acquisition process is prone to the problems of blurred edges, inconspicuous contrast and distortion, which cannot show the defects clearly.To improve image analyzability and clarity, a veneer defect image enhancement method based on AMEF-AGC is proposed herein. First, a veneer defect image is subjected to Gamma correction to obtain multiple underexposed image sequences for which Gaussian and Laplacian pyramids are constructed to determine the weights of the multiple exposure sequence group images.Multiscale fusion is then performed based on these weights. Second, the fused image is converted into the HSV color space, where contrast and brightness enhancements are performed for the luminance component, and then converted back to the RGB color space to obtain an enhanced veneer defect image. Solid wood panels selected herein were pine, poplar and birch with defects including live knots, dead knots, and cracks.Compared with those obtained using several algorithms including AMEF, AGC, improved AGC and GC, this algorithm achieved 6.93% and 5.4% improvements in PSNR and SSIM metrics, respectively.Results demonstrated that the proposed method effectively enhanced veneer defect images with blurred and distorted edges, improved image clarity and visual quality, and made defect parts and details of the image clearer.

A novel FuseDecode Autoencoder for industrial visual inspection: Incremental anomaly detection improvement with gradual transition from unsupervised to mixed-supervision learning with reduced human effort

The industrial implementation of automated visual inspection leveraging deep learning is limited due to the labor-intensive labeling of datasets and the lack of datasets containing images of defects, which is especially the case in high-volume manufacturing with zero defect constraints. In this study, we present the FuseDecode Autoencoder (FuseDecode AE), a novel reconstruction-based anomaly detection model featuring incremental learning. Initially, the FuseDecode AE operates in an unsupervised manner on noisy data containing predominantly normal images and a small number of anomalous images. The predictions generated assist experts in distinguishing between normal and anomalous samples. Later, it adapts to weakly labeled datasets by retraining in a semi-supervised manner on normal data augmented with synthetic anomalies. As more real anomalous samples become available, the model further refines its capabilities through mixed-supervision learning on both normal and anomalous samples. Evaluation on a real industrial dataset of coating defects shows the effectiveness of the incremental learning approach. Furthermore, validation on the publicly accessible MVTec AD dataset demonstrates the FuseDecode AE's superiority over other state-of-the-art reconstruction-based models. These findings underscore its generalizability and effectiveness in automated visual inspection tasks, particularly in industrial settings.

Computer diffraction tomography: a comparative analysis of the use of controlled and wavelet filters for image processing

The paper provides digital processing of 2D X-ray projection images of a Coulomb-type point defect in a Si(111) crystal recorded by a detector against the background of statistical Gaussian noise. A managed filter and a wavelet filter with a 4th-order Daubechies function are used. The efficiency of filtering 2D images is determined by calculating the relative quadratic deviations of the intensities of filtered and reference (noiseless) 2D images averaged over all points. A comparison of the calculated values of the relative deviations of the intensities shows that the considered methods work quite well and both, in principle, can be effectively used in practice for noise processing of X-ray diffraction images used for 3D reconstruction of nanoscale defects of crystal structures.

Defect Detection of Polyethylene Gas Pipeline Based on Convolutional Neural Networks and Image Processing

Abstract In this paper, a method based on image recognition was proposed to detect the defects of polyethylene (PE) gas pipeline, especially the deformation due to the indentation. First, the pipeline -detection VGG (PD-VGG) model was established based on the convolutional neural network (CNN), and appropriate model parameters were optimized through model training. The defect recognition rate of the improved model can reach 94.76%. Following, the weighted average graying algorithm was used to separate the defects characterized by deformation. Then, an improved gamma correction algorithm was applied to achieve image enhancement, and the interference of impurities adhered on intersurface of pipeline was also removed by using multilayer filters. The edge detection of the defect image was completed by using the Canny operator, and following the screening between the target contour and the interference contour by using top-contour. Finally, the algorithm for minimum outer rectangle algorithm was used to fit the defect contour, and the eigenvalues of deformation defects were extracted. The results indicate that the above defect detection method can better extract the deformation contour of the dented pipeline. The high agreement with the experimental results provides a basis for the research of effectively recognizing whether the pipeline has undergone ductile failure only through profile detection of defects.

A Candy Defect Detection Method Based on StyleGAN2 and Improved YOLOv7 for Imbalanced Data.

Quality management in the candy industry is a vital part of food quality management. Defective candies significantly affect subsequent packaging and consumption, impacting the efficiency of candy manufacturers and the consumer experience. However, challenges exist in candy defect detection on food production lines due to the small size of the targets and defects, as well as the difficulty of batch sampling defects from automated production lines. A high-precision candy defect detection method based on deep learning is proposed in this paper. Initially, pseudo-defective candy images are generated based on Style Generative Adversarial Network-v2 (StyleGAN2), thereby enhancing the authenticity of these synthetic defect images. Following the separation of the background based on the color characteristics of the defective candies on the conveyor belt, a GAN is utilized for negative sample data enhancement. This effectively reduces the impact of data imbalance between complete and defective candies on the model's detection performance. Secondly, considering the challenges brought by the small size and random shape of candy defects to target detection, the efficient target detection method YOLOv7 is improved. The Spatial Pyramid Pooling Fast Cross Stage Partial Connection (SPPFCSPC) module, the C3C2 module, and the global attention mechanism are introduced to enhance feature extraction precision. The improved model achieves a 3.0% increase in recognition accuracy and a 3.7% increase in recall rate while supporting real-time recognition scenery. This method not only enhances the efficiency of food quality management but also promotes the application of computer vision and deep learning in industrial production.

Dual-Modal Illumination System for Defect Detection of Aircraft Glass Canopies.

Defect detection in transparent materials typically relies on specific lighting conditions. However, through our work on defect detection for aircraft glass canopies, we found that using a single lighting condition often led to missed or false detections. This limitation arises from the optical properties of transparent materials, where certain defects only become sufficiently visible under specific lighting angles. To address this issue, we developed a dual-modal illumination system that integrates both forward and backward lighting to capture defect images. Additionally, we introduced the first dual-modal dataset for defect detection in aircraft glass canopies. Furthermore, we proposed an attention-based dual-branch modal fusion network (ADMF-Net) to enhance the detection process. Experimental results show that our system and model significantly improve the detection performance, with the dual-modal approach increasing the mAP by 5.6% over the single-modal baseline, achieving a mAP of 98.4%. Our research also provides valuable insights for defect detection in other transparent materials.

SODD-YOLOv8: an insulator defect detection algorithm based on feature enhancement and variable row convolution

Abstract To address the detection challenges in defective images of insulators in transmission lines, including tiny object size, significant scale variations, a wide variety of defects, and complex background interference. In this study, an improved insulator defect detection algorithm is proposed, based on the YOLOv8s framework and combining feature enhancement and deformable convolution techniques. Firstly, to address the image feature distortion problem caused by aerial photography, a deformable convolutional feature extraction module (DCFEM) is introduced, which is designed to enhance the model’s ability to adapt to the local geometric deformation, so as to effectively recover the distorted feature information in the image. Moreover, to enhance the detection ability of the model for small objects, a small object feature enhancement module is designed, which adopts an efficient multi-scale attention mechanism, and aims to enhance the feature extraction ability of small objects, improve the sensitivity to small-size defects, and improve the detection accuracy. Eventually, to optimize the computational efficiency of the model, the average pooling-sparse convolution-batch normalization (BN) module is proposed. This module combines average pooling, sparse convolution and BN techniques to achieve a lightweight model while maintaining a high level of feature extraction capability. Experimental results on the China power line insulator dataset show that the improved model achieves a 4.3 percentage point improvement in the mAP metric compared to YOLOv8s, and the number of parameters in the model is reduced by 10%. The proposed scheme not only improves the accuracy and efficiency of defect detection, but also reduces the demand for computational resources, thus providing a more reliable and efficient solution for insulator defect detection in practical applications.

Analysing the robustness of finger vein recognition: cross-dataset reliability and vein utility

AbstractFinger vein recognition is an emerging biometric trait known for its privacy features. Despite the remarkable performance of deep learning methods like convolutional neural networks on challenging finger vein datasets, their reliability and robustness need further examination. This study evaluates the robustness of three recognition methods—the traditional Miura Method, a supervised convolutional neural network, and an unsupervised convolutional auto-encoder—through the challenging and more realistic scenario of cross-dataset comparisons. We also analyse the reliability of these methods in terms of sample quality. We introduce a novel vein quality metric to measure vein clarity and complexity and compare it against an existing image quality metric, natural image quality evaluator. Our findings reveal differences in how these recognition methods utilise finger vein images for comparisons, highlighting the need for robust recognition techniques in more realistic scenarios. In addition, our vein quality metric effectively detects defective images, reducing the zero false-match rate from 34.98% to 8.18% on the SDUMLA-HMT dataset. These results indicate the need for metrics more focussing on finger vein image characteristics for effective quality assessment for finger vein images.

Research on defect image inversion method based on multiple lift-off values magnetic memory signals

Metal magnetic memory (MMM) technique can detect macroscopic defects and stress concentration zones of ferromagnets, and it can identify the approximate positions of these flaws, while a little further information about defects characteristics can be provided. To promote study in this area, defects were modeled as located magnetic dipoles whose strength and locations should be determined, and the technique of truncated generalized inverse and reduced space inversion were used to image the dipole strength. Through the simulation experiment, we found that the inversion image quality was influenced by the lift-off value. When the magnetic field data for inversion was measured with a small lift-off value, the inner dipoles contributions were easily buried and the inversion image can only shown the surface dipole well. In contrast, if the magnetic field data was measured with a greater lift-off value, it was possible to inverse the deep dipoles while the image was somewhat out of focus. To overcome this limitation, we composed the magnetic field data measured under different lift-off values together for inversion. We used the same model to test this approach and applied it on real crack defects magnetic field data. It demonstrates that the new inversion approach shows better accuracy than the single lift-off value and it was possible to infer the location and size of defect by the evaluation of preset dipoles.

A prior knowledge-embedded deformable convolutional network for vision-based fabric defect detection

The vision-based defect detection of textile surface is an essential problem in evaluating the appearance quality. In previous studies on vision-based defect detection, the following two points were neglected: (1) The proportion of defects on fabric is small, resulting in a low signal-to-noise ratio of defect images and a lack of sampling features; (2) the irregular shape of defects can overlap in position, leading to localization difficulty. In this paper, we propose a prior knowledge-embedded deformable convolutional network (PKE-DCNet) based on deep learning to address these two issues. First, a feature extraction method with prior information is designed with a shape-matching deep clustering module and a region-biased sampling module for detecting defects with complex shape. Then, a defect boundary boxes adaptive generation method is proposed with an anchor-free search mechanism and irregular edge contour computation to detect the key points in the region. Extensive experiments with mixed fabric defect datasets demonstrated that PKE-DCNet reached an overall mAP of 95.36% for six types of defects within a detection speed of 322 FPS, which was better than state-of-the-art methods.

Virtual source total focusing method for surface defects using leaky Rayleigh waves

Leaky Rayleigh waves are sensitive to surface defects and beneficial for automated detection due to their non-contact characteristics. However, the amplitude of the leaky Rayleigh waves is low due to the attenuation of waveform conversion and acoustic waves propagation, which limits the imaging quality for long-distance detection. This study introduces a novel method combining leaky Rayleigh waves detection with virtual source total focusing method. The virtual source (VS) technology enhances the emission energy of leaky Rayleigh waves. Then, the total focusing method (TFM) is applied to acquire high-accuracy images of defects. To reduce noise and artifacts, a weighting function based on the coherence factor (CF) is developed to weight the TFM superimposed signals, thereby achieving high-quality image reconstruction. Experimental results show that compared with the conventional TFM method, the proposed method can effectively improve the amplitude of imaging signals while reducing system noise and imaging artifacts. The lateral accuracy of defects is improved, with the average lateral error of defect size being 0.178 mm. The signal-to-noise ratio (SNR) of ultrasound images is increased by 27.59 dB, and the array performance index (API) of ultrasound images is decreased by 33.78 %. The proposed method provides a new and effective approach for quantitatively assessing the surface defects of metal components using leaky Rayleigh waves.

PE Gas Pipeline Defect Detection Algorithm based on Improved YOLO v5

In order to improve defects detection efficiency in polyethylene (PE) gas pipelines and decrease leakage or other pipelines abnormalities in operation, this research proposed an improved YOLO(You Only Look Once) v5 detection model. First, the collected pipeline defect images were processed in grey scale, which improved the computational efficiency of the computer; then, Gamma transform and double filtering algorithms were applied respectively for image enhancement and noise reduction filtering of defects, which enhanced image quality and reduced image noise. Finally, the improved Sobel algorithm was applied to detect defective image edges and the defects in the image were segmented by adaptive threshold segmentation method to obtain binary images. The obtained binary images were employed to train the improved YOLO v5 detection model. The obtained experimental results showed that, compared with the original algorithm, the improved detection algorithm had better detection efficiency and higher robustness as well as higher recognition for common defects,improved YOLOv5 mAP and recall were 97.18% and 98.03%, respectively, the mAP has increased by 1.33% and the recall has increased by 3.83%,which can achieve the detection and identification of defect types of effects in PE gas pipes.

Detection and classification of surface defects on hot-rolled steel using vision transformers

This study proposes a vision transformer to detect visual defects on steel surfaces. The proposed approach utilizes an open-source image dataset to classify steel surface conditions into six fault categories namely, crazing, inclusion, rolled in, pitted surface, scratches and patches. The defect images are first subject to resizing and then fed into a vision transformer subject to different hyperparameter configurations to determine the most optimal setting to render highest classification performance. The performance of the model is evaluated for different hyperparameter configurations, and the most optimal configuration is examined using the associated confusion matrices. It was observed that the proposed model presents a high overall accuracy of 96.39 % for detection and classification of steel surface faults. The study presents a descriptive insight into the vision transformer architecture and in addition, compares the performance of the current model with the results of other approaches suggested for application in literature. Vision transformers can serve as standalone approaches and suitable alternatives to the widely used convolution neural networks (CNNs) by actuating complex defect detection and classification tasks in real-time, enabling efficient and robust condition monitoring of a wide range of defects.

Learning graph-Fourier spectra of textured surface images for defect localization

In the realm of industrial manufacturing, product inspection remains a significant bottleneck, with only a small fraction of manufactured items undergoing inspection for surface defects. Advances in imaging systems and AI can allow automated full inspection of manufactured surfaces. However, even the most contemporary imaging and machine learning methods perform poorly for detecting defects in images with highly textured backgrounds, that stem from diverse manufacturing processes. This paper introduces an approach based on graph Fourier analysis to automatically identify defective images, as well as crucial graph Fourier coefficients that inform the defects in the images. The approach thereby facilitates precise localization and characterization of defects, amidst highly textured backgrounds. A convolutional neural network model (1D-CNN) was trained with the coefficients of the graph Fourier transform of the images as the input to identify, with classification accuracy of 99.4%, if the image contains a defect. An explainable AI method using SHAP (SHapley Additive exPlanations) was used to further analyze the trained 1D-CNN model to discern important spectral coefficients for each image. This approach sheds light on the crucial contribution of low-frequency graph eigen waveforms to precisely localize surface defects in images, thereby advancing the realization of zero-defect manufacturing.

Entropy Stacked Autoencoder based Diffusion Model (ESADM) and Fuzzy Clustering with Semi Supervised Fuzzy Graph Convolutional Network (FC-SSFGCN) for Rail Surface Defect Detection

Rails, fasteners, and other parts of railway track lines eventually develop flaws due to continuous strain from train operations and direct exposure to the environment; these faults directly affect the safety of train operations. Detecting defects on rail surfaces presents a formidable challenge due to the wide array of possible flaws and their unpredictable nature. However, Defect identification errors, massive variances, inadequate training sample availability, and weak contrast between faults and the surrounding background all contribute to the complexity of this procedure. In this work, rail surface and fastener flaws may be detected non-destructively using a multi-crack detection approach based on the Entropy Stacked Autoencoder Diffusion Model (ESADM). A novel approach, ESAFM, has been developed, combining a rail surface image encoder with a multi-layer, Stacked Autoencoder to extract latent materials from images showcasing different types of cracks. This integration reduces the need for significant processing resources by integrating easily into the conventional physically-based image workflow. Additionally, the Zero Shot-Semi Supervised Fuzzy Class Knowledge Graph (ZS-SSFCKG) method proposes Class Knowledge Graph Construction (CKGC), which constructs a CKG to elucidate the connection between defects and non-defects. Class features are learned by using a Fuzzy Clustering with Semi Supervised Fuzzy Graph Convolutional Network (FC-SSFGCN). The method employs a transformer encoder to capture distant relationships, allowing for the extraction of features from defect samples. This facilitates the acquisition of distinct defect image characteristics, as industrial defects vary in shape and size. The experimental results on the public Railway Track Fault Detection (RTFD)and Rail Surface Defect Datasets (RSDDs) with rail surface defects are collected from rail tracks surface defect detection.

Fabric Defect Detection Using Transfer Learning

Transfer learning in fabric defect detection involves utilizing pre-trained deep learning models on a large dataset, typically from a different domain, and fine-tuning them on a smaller dataset that is specific to fabric defects. By leveraging transfer learning, the limitations of limited annotated data for fabric defect detection can be overcome by utilizing the knowledge gained from a more extensive and diverse dataset. The pre-trained model's learned features are adjusted to recognize specific fabric defect patterns, resulting in more accurate and efficient defect detection. This approach reduces the reliance on a massive labelled dataset for training, which is particularly beneficial in industrial applications where obtaining a vast amount of annotated fabric defect images may be challenging. Ultimately, transfer learning enhances the model's ability to generalize and detect fabric defects with higher precision, thereby contributing to improved quality control in textile manufacturing processes.