检索结果-内蒙古大学图书馆

URS: A Light-Weight segmentation Model for Train Wheelset Monitoring

IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2023年第7期24卷 7707-7716页

作者： Guo, Xiaoxuan Ji, Zhenyan Feng, Qibo Wang, Huihui Yang, Yanyan Li, Zhao Beijing Jiaotong Univ Sch Phys Sci & Engn Beijing 100044 Peoples R China Beijing Jiaotong Univ Sch Software Engn Beijing 100044 Peoples R China Dongguan Nannar Elect Technol Co Ltd Dongguan 523050 Peoples R China St Bonaventure Univ Sch Arts & Sci St Bonaventure NY 14778 USA Zhejiang Univ Coll Comp Sci Hangzhou 310058 Zhejiang Peoples R China

To detect the wheelset deformation and wear, an intuitive method is to first collect wheelset multi-line laser stripe images by monitors, and then extract the centerlines to construct 3D contours that are transferred to the cloud data center by 6G communication. The images, however, contain flairs and fractures due to the influence of environmental interference and reflected light on the smooth surfaces. The image defects affect the accurate extraction of stripe centerlines. To segment the defects and inpaint them, we propose a new lightweight U-shaped segmentation model URS. A target-shaped receptive field is designed to efficiently extract the details of the local, the ring-shaped, and the cross-shaped context around the local, which facilitates segmenting various defects. A scale-select sub-module is designed to adjust the weights of features from different receptive fields. To train the model, a multi-line laser image defect segmentation dataset MLIDSD is constructed. Experiments demonstrate that the performance of our model surpasses twelve SOTA models explicitly and can achieve a balance between the accuracy and the lightweight requirement.

关键词： Convolution Feature extraction Image segmentation Decoding Laser modes Semantics Data mining defect segmentation target-shaped receptive field U-shaped network light-weight multi-line laser images

来源：评论

学校读者我要写书评

暂无评论

The MVTec Anomaly Detection Dataset: A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection

引用

INTERNATIONAL JOURNAL OF COMPUTER VISION 2021年第4期129卷 1038-1059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 detection of anomalous structures in natural image data is of utmost importance for numerous tasks in the field of computer vision. The development of methods for unsupervised anomaly detection requires data on which to train and evaluate new approaches and ideas. We introduce the MVTec anomaly detection dataset containing 5354 high-resolution color images of different object and texture categories. It contains normal, i.e., defect-free images intended for training and images with anomalies intended for testing. The anomalies manifest themselves in the form of over 70 different types of defects such as scratches, dents, contaminations, and various structural changes. In addition, we provide pixel-precise ground truth annotations for all anomalies. We conduct a thorough evaluation of current state-of-the-art unsupervised anomaly detection methods based on deep architectures such as convolutional autoencoders, generative adversarial networks, and feature descriptors using pretrained convolutional neural networks, as well as classical computer vision methods. We highlight the advantages and disadvantages of multiple performance metrics as well as threshold estimation techniques. This benchmark indicates that methods that leverage descriptors of pretrained networks outperform all other approaches and deep-learning-based generative models show considerable room for improvement.

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

来源：评论

学校读者我要写书评

暂无评论

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

引用

IEEE ACCESS 2024年 12卷 44286-44305页

作者： Ku, Donghun Pahk, Heui Jae Seoul Natl Univ Sch Mech & Aerosp Engn Seoul 08826 South Korea

This paper aims to develop a defect-background separated generative adversarial network (GAN) using deep learning and GAN to enhance the accuracy of battery exterior defect inspection. In actual battery production lines, the occurrence rates of defects vary by defect type, making it challenging to create a large, uniform defect dataset due to the time required for defect acquisition. This leads to a reduction in the accuracy of battery exterior defect inspection. To construct a large, uniform defect dataset, this paper proposes a defect-background separated GAN based on the principles of GANs. The defect-background separated GAN performs effective defect and background separation learning by referring to the segmentation labeling of defects. Through dataset augmentation using the defect-background separated GAN, the performance quality of newly generated synthetic defect images has improved, and accuracy in battery exterior defect inspection can be enhanced through extensive dataset training. Experimental results show the lowest Frechet inception distance score among various other methods for the battery exterior defect dataset, and it generates clear synthetic defects perceptible to the human eye. By training defect segmentation on this large, uniform defect dataset, an accuracy of 96.1% and an intersection over union value of 0.71 were achieved. Ultimately, applying this defect inspection network to the actual production line demonstrated a 72% improvement in time efficiency. This demonstrates the stability and robustness of the large, uniform defect dataset generated through the defect-background separated GAN.

关键词： Inspection Generative adversarial networks Deep learning Batteries Training Data models Production defect detection Synthetic data Detection algorithms Battery exterior defect inspection deep learning defect segmentation generative adversarial network synthetic defect

来源：评论

学校读者我要写书评

暂无评论

segmentation of weld defects in radiographs using rank-levelling and background subtraction

引用

INSIGHT 2008年第4期50卷 188-194页

作者： Soo, S. L. Ratnam, M. M. Samad, Z. Khalid, M. A. Univ Sci Malaysia Sch Mech Engn Nibong Tebal 14300 Penang Malaysia Agensi Nuklear Malaysia Intelligent Syst Grp Bangi 43650 Selangor Malaysia

segmentation of defects in a digitised weld radiograph is a precursor to automatic interpretation of the radiograph. Simple thresholding has been shown by several researchers in the past to be unsuitable for isolating defects from the weld bead due to the non-uniform background intensity. The background subtraction method (BSM) using polynomial background estimation was shown by others to be able to produce satisfactory results. However, the BSM using polynomial background estimation is unsuitable when the background is highly non-homogeneous with noise present in the image. In this research, an effective algorithm has been developed to segment the defects automatically from noisy weld radiographs having poor illumination using background subtraction and rank-levelling. The rank-levelling method was used to obtain the background image of the digitised weld radiograph. A comparative study on background estimation by the rank-levelling technique and polynomial surface fitting algorithm shows that background estimation using rank-levelling is able to isolate the defects better compared to polynomial surface fitting algorithm.

关键词： defect segmentation weld radiograph rank-levelling

来源：评论

学校读者我要写书评

暂无评论

AuToDiDAC: Automated Tool for Disease Detection and Assessment for Cacao Black Pod Rot

引用

CROP PROTECTION 2018年 103卷 98-102页

作者： Stanley Tan, Daniel Neil Leong, Robert Franchesca Laguna, Ann Anne Ngo, Courtney Lao, Angelyn Amalin, Divina M. Alvindia, Dionisio G. De La Salle Univ Software Technol Dept Manila Philippines De La Salle Univ Math Dept Manila Philippines De La Salle Univ Comp Technol Dept Manila Philippines De La Salle Univ Biol Dept Manila Philippines De La Salle Univ Ctr Complex & Emerging Technol AdRIC Manila Philippines De La Salle Univ Math & Stat Modeling Unit CENSER Manila Philippines De La Salle Univ Biol Control Res Unit CENSER Manila Philippines Philippine Ctr Postharvest Dev & Mechanizat Nueva Ecija Philippines

Pest control strategies for crop diseases highly depend on visual inspection to assess the severity of the infection, which usually lead to inconsistencies: either over or under assessment. These inconsistencies could be attributed to the limitations of humans to perceive small differences. A more precise disease assessment is needed for better pest management decision, which will result to a more efficient utilization and allocation of resources for farm inputs. This translates to a better income for cacao farmers. This paper introduces a mobile application named AuToDiDAC or Automated Tool for Disease Detection and Assessment for Cacao Black Pod Rot (BPR). AuToDiDAC automatically detects, separates, and assesses the infection level of BPR in cacao through image processing and machine learning techniques. It gives the farmers the capacity to objectively monitor and report the infection level of the BPR compared to the common visual rating for plant disease level of infection. Pixel level accuracy test of the tool showed an average of 84% accuracy on an independent test set of ten cacao pod images.

关键词： Cacao Black pod rot defect segmentation Infection level Disease management tool Decision support

来源：评论

学校读者我要写书评

暂无评论

Fabric defect Detection Based on Annular Gaussian Band-pass Filters

Fabric Defect Detection Based on Annular Gaussian Band-pass ...

引用

5th International Conference on Digital Image Processing (ICDIP)

作者： Han, Runping Su, Jianxia Beijing Inst Fash Technol Sch Informat Engn Beijing 100029 Peoples R China

ISBN: (纸本)9780819493057

A new method for fabric defect detection is proposed. It is based on the filter group containing four annular Gaussian band-pass filters and aimed at detecting defects in fabrics with plain and twill structures. A fabric sample image is processed by using this filter group and the filtered images are obtained. These filtered images are binarized and fused in order to reconstruct the defect binary image that distinguishes defects from texture background. In the performance evaluation and comparison experiments, this method is applied to a variety of fabrics with various defects. Experiment results have confirmed this method has good real time performance and is effective in fabric defect detection.

关键词： Fabric defect detection defect segmentation Gaussian band-pass filter Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Inspection of Wind Turbine Blades Using Image Deblurring and Deep Learning segmentation 18

Inspection of Wind Turbine Blades Using Image Deblurring and...

引用

Conference on Health Monitoring of Structural and Biological Systems XVIII

作者： Lu, Jiale Gao, Qingbin Zhou, Kai Hong Kong Polytech Univ Dept Civil & Environm Engn Hong Kong Peoples R China Harbin Inst Technol Sch Mech Engn & Automat Shenzhen Peoples R China

ISBN: (纸本)9781510672093;9781510672086

Remote and complex work sites of wind turbines limit the accessibility of the condition assessment. Wind turbine blades are subject to sustained wind load and harsh natural environmental conditions, which are vulnerable to various faults. Robotic-enabled sensing technology appears to be a promising solution for an efficient wind turbine blade inspection. Together with the recent advances in image processing and deep learning segmentation, automated inspection of wind turbine blades becomes possible. Nevertheless, it remains a challenging task to quantify the damage accurately due to the complex condition of images concerning motion blurs. To address this issue, an integrated framework, i.e., the combination of a Deblur Generative Adversarial Network v2 (DeblurGAN-v2) and You Only Look Once v8 (YOLO-v8) was proposed in this study. Specifically, the mapping between the motion-blurred images and those in high quality was adopted from the open-access pretrained DeblurGAN-v2, based on which the deblurring performance for wind turbine images with various motion blur scales was discussed concerning the image quality. Subsequently, the transfer learning method was implemented to fine-tune YOLO-v8. The well-trained YOLO v8 was then utilized for target defect segmentation on the deblurred images. Finally, various metrics were calculated to evaluate the segmentation accuracy and efficiency. Results prove a good generalization of DeblurGAN-v2 on wind turbine images and clearly illustrate the enhanced performance of the proposed methodology especially when the motion blur scale is within 35. The integrated framework could serve as a reference for dealing with other fuzzy image-related issues.

关键词： Wind turbine blades Robotic-enabled sensing Image deblurring defect segmentation Deblur Generative Adversarial Network v2 (DeblurGAN-v2) You Only Look Once (YOLO)

来源：评论

学校读者我要写书评

暂无评论

Automatic segmentation of Pores in Weld Images Based on Transition Region Extraction

Automatic Segmentation of Pores in Weld Images Based on Tran...

引用

2nd International Conference on Advanced Design and Manufacturing Engineering (ADME 2012)

作者： Tong, Tong Cai, Yan Sun, Dawei Liu, Peng Shanghai Jiao Tong Univ Shanghai Key Lab Mat Laser Proc & Modificat Shanghai 200240 Peoples R China

ISBN: (纸本)9783037855027

In allusion to the complex images of weld defects, weak contrast between the target and the background, a new segmentation method based on gray level difference transition region extraction is proposed. The paper analyzes the characteristic of weld defects, and then low-pass filtering and contrast enhanced are used to enhance the clarity. Finally, we extract the transition region and confirm a threshold for defects segmentation. The experimental results show that the method can extract the transition region more accurate, and segment the image much better in complex environment.

关键词： Contrast highlighting Transition region defect segmentation Thresholding

来源：评论

学校读者我要写书评

暂无评论

Automatic Labeling in Image segmentation and Classification for TFT-LCD Manufacturing 19

Automatic Labeling in Image Segmentation and Classification ...

引用

19th IEEE International Conference on Mechatronics and Automation (IEEE ICMA)

作者： Liu, Chang Vaassen, Sebastian Manoj, Lakshmi Zhan, Xiao Qing Xu, Chen Ajay, Someshwar Rudra Lu, Zi Yan Wittstamm, Max Jain, Samarth Zhang, Chao Drescher, Benny Hong Kong Ind Artificial Intelligence Robot Ctr Unit 1207 1211 Shatin 12-F Bldg 19WHong Kong Sci Pk Hong Kong Peoples R China Ctr Connected Ind RWTH Aachen CampusCampus Blvd 55 D-52074 Aachen Germany TCL Corp Res Hong Kong Co Ltd Shatin 5-F Bldg 22E22 Sci Pk East Ave Hong Kong Peoples R China

ISBN: (数字)9781665408530

ISBN: (纸本)9781665408530;9781665408523

Product quality inspection of Thin-film transistor-liquid crystal display (TFT-LCD) is time-consuming and labor-intensive. An automatic algorithm-based defect inspection system solves these problems by reducing the time and manual labor involved. This research work proposes an AI-based LCD inspection vision system that evaluates defects by AI-based defect segmentation and classification. Time-consuming pixel-level labeling of defects can be eliminated by applying weakly-supervised learning methods. The efforts of quality control personnel for model training can be reduced which is especially important in high-mix production with a high amount of change-overs and production process adjustments. Therefore, PP-CAM (Precise-Puzzle-CAM) based defect segmentation method is proposed to cope with diverse TFT-LCD defect shapes and sizes. Secondly, a region of interest-supported classification method is developed to enable cropping of small-scale TFT-LCD. The performance of the AI methods are investigated using industrial TFT-LCD manufacturing datasets of two manufacturing processes.

关键词： LCD Vision inspection defect segmentation defect classification Product quality evaluation

来源：评论

学校读者我要写书评

暂无评论

Fabric defects segmentation approach based on texture primitive

Fabric defects segmentation approach based on texture primit...

引用

6th International Conference on Machine Learning and Cybernetics

作者： Zhu, Shuang-Wu Hao, Hong-Yang Li, Peng-Yang Shi, Mei-Hong Qi, Hua Northwestern Polytech Univ Coll Elect Informat Xian 710072 Peoples R China Xian Polytech Univ Coll Text & Mat Xian 710048 Peoples R China

ISBN: (纸本)9781424409723

A new fabric defects segmentation approach based on texture primitive is put forward in the paper, which consists of four steps:1) calculating texture primitive template using auto-correlation function;2) enhancing defect image through calculating difference between each texture primitive and primitive template;3) constructing mean gradation image to attenuate the high frequent background information;4) segmenting defect images according to threshold values acquired automatically through Otsu's approach, etc. Validity and robustness of the approach were proved by different fabric defect images segmentation experiment.

关键词： texture primitive fabric detection image enhancement defect segmentation

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：