检索结果-内蒙古大学图书馆

A Semantic Information Decomposition Network for Accurate segmentation of Texture defects

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS 2023年第7期19卷 8319-8327页

作者： Yang, Hua Hu, Jiale Yin, Zhouping Wang, Zhengjia Huazhong Univ Sci & Technol Sch Mech Sci & Engn State Key Lab Digital Mfg Equipment & Technol Wuhan 430074 Peoples R China Hubei Univ Technol Sch Mech Engn Hubei Key Lab Modern Mfg Quant Engn Wuhan 430068 Peoples R China

defect detection on textured surfaces remains a challenging task due to the wide range of textures and defects. Current unsupervised learning-based texture defect detection methods based on texture background reconstruction cannot detect texture defects with high precision because it is difficult to guarantee a high-precision reconstruction of the texture background while suppressing the defect foreground. In this study, we propose a novel semantic information decomposition network (SIDN) for accurate texture defect segmentation. The SIDN is trained on artificial defective images produced by a defect generation module (DGM). First, the SIDN uses a feature extraction module (FEM) to extract latent features with both texture semantic information and defect semantic information. Then, a novel feature separation extraction module (FSEM) for decomposing the texture semantic information and defect semantic information from the feature map generated by the FEM is proposed, preventing the coupling of the texture and defect semantic information from affecting the final segmentation accuracy. Next, a novel global semantic relation module (GSRM) is proposed to determine the relevance of the global semantic information to comprehensively consider the context and improve the feature representation. Finally, a segmentation module (SM) that directly segments the textures and defects instead of reconstructing the texture background is proposed. The final detection result is obtained by calculating a weighted average of the texture and defect segmentation results. The extensive experimental tests with the most popular and most challenging texture defect dataset demonstrate that the SIDN achieves accurate segmentation of various texture defects without using real defect samples.

关键词： Anomaly detection defect segmentation feature clustering feature separation texture defect detection

来源：评论

学校读者我要写书评

暂无评论

A Pixel-Wise segmentation Method for Automatic X-Ray Image Detection of Chip Packaging defects

引用

IEEE TRANSACTIONS ON COMPONENTS PACKAGING AND MANUFACTURING TECHNOLOGY 2024年第8期14卷 1520-1527页

作者： Wang, Jie Li, Gaomin Zhou, Yuezheng Bai, Haoyu Li, Xuan Zhong, Lijun Zhang, Xiaohu Sun Yat Sen Univ Sch Aeronaut & Astronaut Guangzhou 510725 Peoples R China China Aerosp Components Engn Ctr Beijing 100094 Peoples R China

Integrated circuit chips are the most common electronic components, and visual internal defect detection is essential for ensuring product quality following packaging. However, efficient detection of chip internal defects is challenging due to the complexity of the background and the faintness of defects. To overcome the above difficulties, a novel deep learning-based defect segmentation framework is proposed, which relies on an image preprocessing (IPP) scheme and a defect segmentation network (DSNet). The IPP is composed of a rotation correction algorithm and a region segmentation algorithm for removing the influence of background and obtaining chip packaging region. The DSNet is proposed to precisely and efficiently segment the internal defects. To address the scarcity of data and avoid overfitting, we proposed the lightweight convolution block by using depth-wise separable convolution (DWSC) to reduce the number of parameters. Besides, the attention gate (AG) module is incorporated into the skip connection to handle the shape varieties of the defects. Moreover, the Focal loss function is designed to guide the network to pay attention to small defects that are difficult to distinguish. The robustness and adaptability of the proposed method are evaluated on three typical types of chip X-ray datasets from real-world inspection lines. Experimental results show that the proposed framework achieves a satisfactory tradeoff between detection accuracy and speed with an $F1$ -score of 72.69% and a frames per second (FPS) of 17.5 on average, resulting in superior segmentation performance even with sparse or insufficient data.

关键词： X-ray imaging Image segmentation Packaging Inspection Feature extraction defect detection Object segmentation Attention gate (AG) chip packaging defect defect segmentation lightweight convolution X-ray

来源：评论

学校读者我要写书评

暂无评论

Surface defects Detection and Identification of Lithium Battery Pole Piece Based on Multi-Feature Fusion and PSO-SVM

引用

IEEE ACCESS 2021年 9卷 85232-85239页

作者： Xu, Changlu Li, Linsheng Li, Jiwei Wen, Chuanbo Shanghai Dianji Univ Elect Inst Shanghai 201306 Peoples R China

In order to realize the automatic detection of surface defects of lithium battery pole piece, a method for detection and identification of surface defects of lithium battery pole piece based on multi-feature fusion and PSO-SVM was proposed in this paper. Firstly, image subtraction and contrast adjustment were used to preprocess the defect image to weaken the influence of non-defective areas and enhance the defect features. Then, Canny algorithm and the AND logical operation were used to extract the image of defect area. Next, the texture feature, edge feature, and HOG feature were combined to extract the feature of the defect area image. Finally, the support vector machine (SVM) optimized by particle swarm optimization (PSO) was used to automatically identify and classify defect images. The experimental results show that the proposed method in this paper can effectively detect surface multiple types defects of lithium battery pole piece, and the average recognition rate of defects reaches 98.3%, which is an effective and feasible automatic defect detection and identification method.

关键词： Lithium batteries Feature extraction Image edge detection Surface treatment Surface morphology Support vector machines Classification algorithms Machine vision surface defects of lithium battery pole piece defect detection support vector machine defect segmentation

来源：评论

学校读者我要写书评

暂无评论

Pixel-accurate road crack detection in presence of inaccurate annotations

引用

NEUROCOMPUTING 2022年第0期480卷 1-13页

作者： Rill-Garcia, Rodrigo Dokladalova, Eva Dokladal, Petr Univ Gustave Eiffel ESIEE Paris CNRS LIGM F-77454 Marne La Vallee France Univ Gustave Eiffel Navier Lab CNRS Ecole Ponts ParisTech F-77454 Marne La Vallee France MINES Paris PSL Res Univ Ctr Math Morphol 35 Rue St Honore F-77305 Fontainebleau France

Recent road crack detection methods obtain appealing scores but typically allow a few pixel tolerance margin. This is acceptable for locating cracks, but not for measuring their width (indicator of the cracks' severity). Our baseline model, U-VGG19, obtains an F-score of 71.77% on CrackForest, which is superior to other approaches when no tolerance is admitted. However, increasing the scores without tolerance is difficult due to inaccurate annotations. We propose a novel synthetic dataset, Syncrack, as a benchmark for the evaluation of training with inaccurate annotations. Our results show that inaccurate annotations have a detrimental impact on the F-measure, decreasing it by up to 20%. To overcome this, we study label noise correction techniques using weakly supervised learning. Training U-VGG19 with these corrected labels improves the results on Syncrack by up to 12%. Obtained results on the CrackForest and Aigle-RN datasets support that these approaches are useful for real-life data too. (c) 2022 Elsevier B.V. All rights reserved.

关键词： Crack detection Deep learning defect segmentation Inaccurate labels Weakly supervised learning

来源：评论

学校读者我要写书评

暂无评论

Beyond Dents and Scratches: Logical Constraints in Unsupervised Anomaly Detection and Localization

引用

INTERNATIONAL JOURNAL OF COMPUTER VISION 2022年第4期130卷 947-969页

作者： Bergmann, Paul Batzner, Kilian Fauser, Michael Sattlegger, David Steger, Carsten MVTec Software GmbH Arnulfstr 205 D-80634 Munich Germany Tech Univ Munich Dept Informat Boltzmannstr 3 D-85748 Garching Germany

The unsupervised detection and localization of anomalies in natural images is an intriguing and challenging problem. Anomalies manifest themselves in very different ways and an ideal benchmark dataset for this task should contain representative examples for all of them. We find that existing datasets are biased towards local structural anomalies such as scratches, dents, or contaminations. In particular, they lack anomalies in the form of violations of logical constraints, e.g., permissible objects occurring in invalid locations. We contribute a new dataset based on industrial inspection scenarios that evenly covers both types of anomalies. We provide pixel-precise ground truth data for each anomalous region and define a generalized evaluation metric that addresses localization ambiguities that can arise for logical anomalies. Furthermore, we propose a novel algorithm that improves over the state of the art in the joint detection of structural and logical anomalies. It consists of a local and a global network branch. The first one inspects confined regions independent of their spatial locations in the input image and is primarily responsible for the detection of entirely new local structures. The second one learns a globally consistent representation of the training data through a bottleneck that enables the detection of violations of long-range dependencies, a key characteristic of many logical anomalies. We perform extensive evaluations on our new dataset to corroborate our claims.

关键词： Anomaly detection Novelty detection Datasets Unsupervised learning defect segmentation Performance metrics

来源：评论

学校读者我要写书评

暂无评论

Window Feature-Based Two-Stage defect Identification Using Magnetic Flux Leakage Measurements

引用

IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT 2018年第1期67卷 12-23页

作者： Liu, Jinhai Fu, Mingrui Liu, Feilong Feng, Jian Cui, Kuangqing Northeastern Univ Coll Informat Sci & Engn Shenyang 110004 Liaoning Peoples R China CNOOC EnerTech Equipment Technol Co Ltd Tianjin 100024 Peoples R China

Magnetic flux leakage (MFL) testing, one of the nondestructive testing methods, is widely adapted by approximately 90% of in-service pipelines. It is very important to identify defects in MFL testing. This paper presents a method to detect defects and determine precise defect regions using window features generated from MFL measurements. The main contributions are: 1) four novel window features of defects, namely, saliency, contrast, center point (CP), and fingerprint with parameters, are presented, which provide sufficient information for defect identification. In particular, a Bayesian method is developed to estimate the near-optimal parameters of these features and 2) a novel two-stage identification process is proposed, which can not only detect defects, but also segment the complicated defect regions precisely. The performance of the proposed method is demonstrated by real MFL measurements collected from experimental pipelines and in-service pipelines, respectively. The results show that the proposed method has satisfied accuracy for defect identification of experimental and engineering application.

关键词： Boundary determination defect identification defect segmentation magnetic flux leakage (MFL)

来源：评论

学校读者我要写书评

暂无评论

An industrial defect detection algorithm based on CPU-GPU parallel call

引用

MULTIMEDIA TOOLS AND APPLICATIONS 2023年第28期82卷 44191-44207页

作者： Li, Zhu Lin, Hong-wei Liu, Yuan-yuan Chen, Chong Xia, Yun-fei Hangzhou Dianzi Univ Sch Elect & Informat Hangzhou 310000 Peoples R China

The workpiece positioning and defect segmentation are two key steps in the workpiece detection process. This paper has designed a CPU-GPU parallel call algorithm based on real industrial quality inspection conditions to realize the high-speed workpiece defect detection. The algorithm can fully utilize all computing resources of the industrial control computer while simultaneously accomplishing the workpiece positioning and defect segmentation tasks. Moreover, to reduce the workpiece defect detection's scope and improve the defect segmentation algorithm's efficiency, the proposed method uses the workpiece positioning results. As for the positioning task, we have designed the double pyramid method to enhance the positioning speed. When it comes to the defect segmentation task, we have introduced the lightweight network to improve the workpiece segmentation speed. Considering that the current general data sets are of the workpiece local image(s) post cutting, we set up a new data set to reflect the situation in the industrial field. It consists of images taken from real industrial fields that can better verify the whole quality inspection algorithm process, including the positioning and segmentation algorithms. According to our experiment, our algorithm accomplished the positioning and defect segmentation tasks at a speed of 116FPS. Additionally, the segmentation accuracy reached 75.12% Mean IoU.

关键词： Workpiece positioning defect segmentation CPU-GPU parallel call algorithm Double pyramid method Lightweight network

来源：评论

学校读者我要写书评

暂无评论

Attention-Guided Multitask Learning for Surface defect Identification

引用

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS 2023年第9期19卷 9713-9721页

作者： Sampath, Vignesh Maurtua, Inaki Martin, Juan Jose Aguilar Rivera, Andoni Molina, Jorge Gutierrez, Aitor Tekniker Autonomous & Intelligent Syst Unit Gipuzkoa 20600 Spain Univ Zaragoza Dept Design & Mfg Engn Zaragoza 50009 Spain

Surface defect identification is an essential task in the industrial quality control process, in which visual checks are conducted on a manufactured product to ensure that it meets quality standards. The convolutional neural network (CNN)-based surface defect identification method has proven to outperform traditional image processing techniques. However, the real-world surface defect datasets are limited in size due to the expensive data generation process and the rare occurrence of defects. To address this issue, this article presents a method for exploiting auxiliary information beyond the primary labels to improve the generalization ability of surface defect identification tasks. Considering the correlation between pixel-level segmentation masks, object-level bounding boxes, and global image-level classification labels, we argue that jointly learning features of the related tasks can improve the performance of surface defect identification tasks. This article proposes a framework named defect-Aux-Net, based on multitask learning with attention mechanisms that exploit the rich additional information from related tasks with the goal of simultaneously improving robustness and accuracy of the CNN-based surface defect identification. We conducted a series of experiments with the proposed framework. The experimental results showed that the proposed method can significantly improve the performance of state-of-the-art models while achieving an overall accuracy of 97.1%, Dice score of 0.926, and mean average precision of 0.762 on defect classification, segmentation, and detection tasks.

关键词： Deep learning defect classification defect detection defect segmentation machine vision multitask learning (MTL) quality control surface defect detection

来源：评论

学校读者我要写书评

暂无评论

defect identification in magnetic tile images using an improved nonlinear diffusion method

引用

TRANSACTIONS OF THE INSTITUTE OF MEASUREMENT AND CONTROL 2021年第11期43卷 2413-2424页

作者： Ben Gharsallah, Mohamed Ben Braiek, Ezzedine Natl Higher Sch Engn ENSIT Res CEREP Unit 5 Ave Taha Hussein Tunis 1008 Tunisia

Visual inspection of surface defects is a crucial step in the magnetic tile manufacturing process. Magnetic tile images suffer from a non-uniform illumination, texture and noise that disperse irregularly in flawless image areas. As a result, common edge detection and threshold segmentation techniques fail to identify these kinds of defects. In this work, we present a robust algorithm for defect identification in magnetic tile images. The proposed method is based on a new anisotropic diffusion filtering model. Unlike traditional anisotropic diffusion models that take into account only gradient magnitude information, the proposed model combines together gradient magnitude and a new local difference image feature. The aim is to remove bright shapes and undesirable artifacts in the faultless region in magnetic tile images. In addition, the method activates a smoothing process in the flawless region to homogenize the background and simultaneously a sharpening in the defect boundaries to highlight anomalies. Experimental results on a number of magnetic tiles samples containing different types of defects have demonstrated the efficiency of the proposed diffusion method.

关键词： Magnetic tile images defect segmentation anisotropic diffusion image filtering integral images

来源：评论

学校读者我要写书评

暂无评论

Surface splicing defect analysis and application of polarization maintaining fiber using graph cut with illumination priors

引用

INFRARED PHYSICS & TECHNOLOGY 2014年 66卷 125-135页

作者： Liu, Haoting Wang, Wei Gao, Feng Li, Jie Chen, Kai China Acad Aerosp Elect Technol Beijing 100094 Peoples R China Astronaut Res & Training Ctr China Beijing 100094 Peoples R China Satellite Maritime Tracking & Control Dept China Jiangyin 214431 Peoples R China

A new surface splicing defect analysis and application technique of Polarization Maintaining (PM) fiber is proposed. In contrast to the traditional artificial experience based analysis method, we not only develop an automatic defect segmentation technique for the fiber splicer but utilize the image features of splicing defect to assist evaluation of the splicing effect. First, we employ a standard fiber splicer to implement the splicing operation. Both the visible and the hot (infrared) images are captured during that processes. Second, we use the image processing techniques to analyze the image features for both the visible and the hot images. The Hough line detection is used to monitor the core-offset or the angle tilt problems of spliced fibers in visible image. A new Graph Cut Model (GCM), which uses the Multivariate Gaussian Mixture Model (MGMM) as the illumination prior of transmitted rays, is employed to segment the splicing defect in hot image. Third, multiple defect image features, such as the linear edge, the defect shape, and the inertia moment are all computed for the description of defective region. Finally, the SVM classifier is employed to evaluate the fiber splicing effect. The defect features, the splice loss, the extinction ratio, together with the final precision output of optical component are utilized to train the SVM. By using this method, a reliable quality control measurement for the optical component of aerospace optoelectronic apparatus is developed. Many experiments have verified the validity of the proposed method. (C) 2014 Elsevier B.V. All rights reserved.

关键词： Fiber splicing defect segmentation Graph cut Illumination priors Splicing evaluation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：