检索结果-内蒙古大学图书馆

2022 IEEE International Conference on High Voltage Engineering and Applications, ICHVE 2022

作者： Fu, Yuwei Liang, Liejuan Zhang, Guowei Zhang, Zhiyu Chen, Chi Wang, Chuang Xi'an University of Technology Xi'an China Xi'an Xd High Voltage Apparatus Co. Ltd Xi'an710018 China

ISBN: (数字)9781665407502

ISBN: (纸本)9781665407502

With the development of sensing technology, a large number of partial discharge (PD) time domain data are generated in the field of gas-insulated integrated electrical appliances (GIS). Traditional pattern recognition methods cannot identify the defect types of such data. At the same time, deep learning methods generally face the problem of small samples. In order to solve the above problems, this paper proposes a PD pattern recognition method based on transfer learning and DenseNet model. Firstly, the finite-difference-time-domain(FDTD) method is used to simulate PD in GIS, and the time domain waveform image dataset of four PD defects was established. Secondly, image processing techniques such as image enhancement, normalization and image compression are used to process the time-domain waveform spectrum. Finally, Using CNN network and transfer learning, the recognition accuracy of the model is improved to 96.38%, and the convergence is obviously improved. The identification performance of different CNN network structures is studied, and the results show that the accuracy of DenseNet model is significantly higher than other structures. This study can be used to diagnose PD defects in GIS equipment, © 2022 IEEE.

关键词： Partial discharges

来源：评论

学校读者我要写书评

暂无评论

Morphological Segmentation with Modified Otsu’s Method and Multi-Scale Morphological processing for Shape recognition

Morphological Segmentation with Modified Otsu’s Method and ...

引用

Innovations in High Speed Communication and Signal processing (IHCSP), International Conference on

作者： Manish Nagpal N.T. Velusudha Deepak Minhas Supriya Rai Ashwini Ranjit Nawadkar D. Bhanu Centre of Research Impact and Outcome Chitkara University Punjab India Department of ECE Prince Shri Venkateshwara Padmavathy Engineering College Chennai India Chitkara Centre for Research and Development Chitkara University Himachal Pradesh India School of Mangement UG ManagementJAIN (Deemed to be University) Bangalore Karnataka India AI & Ds Vishwakarma Institute of Technology Pune India Department of Information Technology Karpagam Institute of Technology Coimbatore India

ISBN: (数字)9798350368949

ISBN: (纸本)9798350368956

Motological segmentation is essential in the domains of medical imaging, object recognition, and remote sensing. Classical approaches, such as Otsu's thresholding, fail miserably in the presence of noise, light, and complex backgrounds. These concerns are addressed by creating Modified Otsu's Method (MOM). Slow and precise segmentation, adaptive thresholding, and morphological processes are two new areas that MOM aims to tackle. The proposed MOM makes both local and global adjustments taking into account the image and text in order to mitigate the impact of variations in light intensity and noise level. Accurate segmentation and less noise artefacts are achieved by the use of dilation and erosion. On datasets including medical, nature, and remote sensing images, the suggested MOM algorithm achieves better results than previous top-tier segmentation approaches as well as Otsu's approach to generic thresholding based on experimental quantities. This study investigates the optimal execution time of the method for image processing in real time in terms of computing cost. Automated safety devices and medical imaging analysis are two examples of applications where MOM stands out for its exceptional performance in picture segmentation. Therefore, in order to make MOM more versatile and effective in a wide range of imaging contexts, future studies will use machine learning methods.

关键词： image segmentation Accuracy Thresholding (Imaging) Noise Real-time systems Object recognition Method of moments remote sensing Biomedical imaging Noise level

来源：评论

学校读者我要写书评

暂无评论

Parallel Optimization of remote sensing image Geometric Correction on DCU

Parallel Optimization of Remote Sensing Image Geometric Corr...

引用

International Conference on Computer Engineering and Applications (ICCEA)

作者： Runjie Liu Siqi Wang National Supercomputing Center in Zhengzhou Zhengzhou University Zhengzhou China School of Computer and Artificial Intelligence Zhengzhou University Zhengzhou China

ISBN: (数字)9798350386776

ISBN: (纸本)9798350386783

Geometric correction is essential in the preprocessing of remote sensing images, ensuring the precision and reliability of spatial data, crucial for further analysis such as target recognition and feature extraction. Nevertheless, handling a lot of high-resolution images often encounters computational inefficiencies and limited real-time processing capabilities. This paper introduces a parallel optimization approach for geometric correction on multi-DCU heterogeneous clusters, leveraging a detailed thread mapping strategy tailored to the DCU architecture. It capitalizes on the specific hardware characteristics and execution models of the DCU for parallel processing. The study delves into the common bicubic interpolation resampling method used in geometric correction, optimizing thread mapping to match the DCU's computational and storage architecture. This method allows a single thread to handle multiple pixel computations, enhancing data reuse during interpolation. The research accomplishes parallel geometric correction processing on DCU and significantly curtails memory access through optimization, improving algorithmic efficiency. Comparative experimental outcomes show that the proposed parallel strategy markedly outperforms traditional serial CPU processing, multicore CPU processing, and unoptimized DCU acceleration, evidencing substantial speedup across processing different image sizes.

关键词： Interpolation Multicore processing Target recognition Instruction sets Computational modeling Parallel processing Spatial databases

来源：评论

学校读者我要写书评

暂无评论

Ship Detection in Optical sensing images Based on Yolov5 12

Ship Detection in Optical Sensing Images Based on Yolov5

引用

12th International Conference on Graphics and image processing (ICGIP)

作者： Chen, Yuwen Zhang, Chao Qiao, Tengfei Xiong, Jianlin Liu, Bin Beijing Res Inst Telemetry Beijing 100094 Peoples R China

ISBN: (纸本)9781510642782;9781510642775

Ship detection technology is an important development direction in the field of optical remote sensing image processing. In recent years, convolutional neural networks have achieved good results in ship target detection and recognition. We train the latest model YOLOv5 on our dataset in this paper. The results show that YOLOv5 can be well applied in the field of ship detection.

关键词： Ship detection optical remote sensing image processing YOLOv5

来源：评论

学校读者我要写书评

暂无评论

Current Advances in Hyperspectral remote sensing in Urban Planning 3

Current Advances in Hyperspectral Remote Sensing in Urban Pl...

引用

3rd International Conference on Intelligent Computing, Instrumentation and Control Technologies, ICICICT 2022

作者： Nisha, A. Anitha, A. Noorul Islam Centre for Higher Education Department of Computer Application Kerala 629 180 India

ISBN: (数字)9781665410052

ISBN: (纸本)9781665410052

Hyperspectral cameras collect loads of spectral bands of notably slender bandwidths simultaneously thereby to extract the spectral signatures of items and substances which offers a better recognition and identification of items on city areas. In this paper the precise specific application of hyperspectral remote sensing to city planning is examined. In recent times more than 55% of the population can stay in towns and the non-stop increase in urban vicinity causes severe damage to biodiversity such as pollution, global warming and many more. Consequently city planning has turned out to be vital to make life secure in towns. Hyperspectral imaging is beneficial to detect, discriminate, and recognize city buildings, sealed soils, vegetation, water surfaces, roads, etc and to assess their real locations and areas to create a better environment. This paper additionally gives an outline of hyperspectral image processing strategies and suggestions for proper urban making plans using hyperspectral technology. © 2022 IEEE.

关键词： remote sensing

来源：评论

学校读者我要写书评

暂无评论

SAR image change detection method based on PPNN

引用

Science China(Information Sciences) 2021年第8期64卷 257-259页

作者： Guoli NIE Guisheng LIAO Cao ZENG National Laboratory of Radar Signal Processing Xidian University International Coopreation Base of Integrated Electronic Information System of Ministry of Science and Technology Xidian University 2011 Collaborative Innovation Center of Information Sensing Xidian University Radar Cognitive Detection Imaging and Target Recognition Overseas Expertise Introduction Center for Discipline Innovation (“111 Center”)Xidian University

Dear editor,In remote sensing research, change detection is said to be a hot spot because of its many applications [1], and lots of synthetic aperture radar(SAR) image change detection methods are proposed. In this study, to address the problem that the accuracy of SAR image change detection needs to be improved, a multi-feature parallel probabilistic neural network(PPNN) is proposed.

关键词： PNN SAR image change detection method based on PPNN image

来源：评论

学校读者我要写书评

暂无评论

remote sensing Based Advance image Captioning Improved Feature Attention

Remote Sensing Based Advance Image Captioning Improved Featu...

引用

pattern recognition and Artificial Intelligence (PRAI), International Conference on

作者： Ammara Sattar Muhammad Assam Tahani Jaser Alahmadi Uzair Aslam Bhatti Hao Tang Muhammad Aamir Harbin Engineering University Harbin China Department of Software Engineering University of Science and Technology Bannu KPK Pakistan College of Computer and Information Sciences Princess Nourah bint Abdulrahman University Riyadh Saudi Arabia School of information and communication engineering Hainan University Haikou China College of Computer Science Huanggang Normal University Huanggang China

ISBN: (数字)9798350350890

ISBN: (纸本)9798350350906

In the domain of image description generation, the focal point mechanism is crucial for highlighting the most relevant features within an image, thereby influencing the quality of generated descriptions. To address this challenge, we introduce an innovative optimization module called the Enhanced Focal (EF) mechanism, which builds upon the traditional multi-head focal point mechanism. The EF module is designed to assess the correlation levels between objects within an image, thereby guiding the process of subtitle generation more effectively. The EF mechanism integrates the output of the multi-head focal point mechanism with the current context (i.e., the query) to generate two components: an “information vector” and an “attention gate.” The “attention gate” is then applied to the “information vector” to create a refined focal point mechanism. This refined mechanism undergoes element-wise multiplication with the “information vector,” producing attention information that reveals a strong correlation between the target and the background in the image. This process leverages the Long Short-Term Memory (LSTM) mechanism to enhance the model's capability to capture intricate relationships. Upon training and evaluating our enhanced model on the MS COCO dataset, we observed significant improvements in key evaluation metrics. The BLEU-1 and METEOR scores reached 73.01% and 28.81%, respectively, showcasing the effectiveness of the proposed EF mechanism in improving image description generation.

关键词： Training Correlation Accuracy Logic gates Vectors Stability analysis pattern recognition Long short term memory Optimization remote sensing

来源：评论

学校读者我要写书评

暂无评论

4th International Conference on image processing and Capsule Networks, ICIPCN 2023

4th International Conference on Image Processing and Capsule...

引用

4th International Conference on image processing and Capsule Networks, ICIPCN 2023

ISBN: (纸本)9789819970926

The proceedings contain 48 papers. The special focus in this conference is on image processing and Capsule Networks. The topics include: Multispectral Fusion of Multisensor image Data Using PCNN for Performance Evaluation in Sensor Networks;enhanced Feature Fusion from Dual Attention Paths Using Feature Gating Mechanism for Scene Categorization of Aerial images;Histopathology Breast Cancer Classification Using CNN;Brain Tumor recognition from MRI Using Deep Learning with Data Balancing Methods and Its Explainability with AI;Multi-class Plant Leaf Disease Classification on Real-Time images Using YOLO V7;semantic image Segmentation of Agricultural Field Problem Areas Using Deep Neural Networks Based on the DeepLabV3 Model;u-Net-Based Segmentation of Coronary Arteries in Invasive Coronary Angiography;modern Challenges and Limitations in Medical Science Using Capsule Networks: A Comprehensive Review;securing Data in the Cloud: The Application of Fuzzy Identity Biometric Encryption for Enhanced Privacy and Authentication;Modified U-Net and CRF for image Segmentation of Crop images;flameGuard: A Smart System for Forest Fire Detection and Control;classification and Analysis of Chilli Plant Disease Detection Using Convolution Neural Networks;a New Multi-level Hazy image and Video Dataset for Benchmark of Dehazing Methods;change Detection for Multispectral remote sensing images Using Deep Learning;studies on Movie Soundtracks Over the Last Five Years;an Enhanced Real-Time System for Wrong-Way and Over Speed Violation Detection Using Deep Learning;tea Leaf Disease Classification Using an Encoder-Decoder Convolutional Neural Network with Skip Connections;EEG Signal Feature Extraction Using Principal Component Analysis and Power Spectral Entropy for Multiclass Emotion Prediction;preface.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Optimization of Edge Detection Algorithm Based on Binocular Camera

Research on Optimization of Edge Detection Algorithm Based o...

引用

Intelligent Communications and Computing (ICC), International Conference on

作者： Xiaoxu Zhang Sen Wang Guotao Cong Weiwei Zou College of Intelligent and Electronic Engineering Dalian Neusoft University of Information Dalian China

ISBN: (数字)9798331530365

ISBN: (纸本)9798331530372

Edge detection is one of the core technologies in digital image processing and computer vision, and has been widely applied in fields such as image recognition and remote sensing image processing. This article proposes an optimization scheme for edge detection algorithm, which collects image information through binocular cameras and caches it in memory. After grayscale transformation preprocessing, the classic SOBEL algorithm and optimized SOBEL algorithm are used to extract edge information, obtain binary images, and display them on the LCD screen in real time. Through practical testing, the optimized algorithm can effectively solve the problem of detecting edges that are too coarse in the original algorithm, and improve the accuracy of detection and positioning. This system can be widely applied in fields such as medical imaging and autonomous driving.

关键词： image edge detection Gray-scale Cameras Real-time systems Liquid crystal displays Optimization Field programmable gate arrays remote sensing Biomedical imaging Testing

来源：评论

学校读者我要写书评

暂无评论

Vector processing in NUMA Systems

Vector Processing in NUMA Systems

引用

作者： Hu, Zhewen Delft University of Technology

学位级别：硕士

Over the last decade, applications like self-driving, image recognition and speech processing are having more and more impact on the society, all these applications are based on machine learning, and machine learning is all about metrics and vectors. For that reason, vector processors are getting attraction again. Most of the previous research of vector processor focuses on single-core performance, and most of today's large-scale computer system use non-uniform memory access (NUMA) architecture, how to efficiently deploy the vector processor in a NUMA environment remains a problem. NUMA is a shared memory model, with a NUMA system, there are multiple memories distributed in the system and usually each NUMA node has one memory. It will take the processors longer to access memories in other nodes then the memory in the same node, and this feature shows some opportunities to increase the performance of the NUMA system by accelerating the remote memory accesses. In this thesis, a subset of the PARSEC benchmark are vectorized for both ARM Scalable Vector Instructions (SVE) and RISC-V Vector Instructions (RVV), and the single-core performance of these two types of processors will be compared based on this benchmark. Then a NUMA system is made for ARM SVE and the memory access pattern is analysed with gem5 simulator.

关键词：

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：