检索结果-内蒙古大学图书馆

5th Chinese conference on pattern recognition and Computer Vision (PRCV)

作者： Ji, Haowei Di, Huijun Wang, Shunzhou Shi, Qingxuan Beijing Inst Technol Sch Comp Sci & Technol Beijing Lab Intelligent Informat Technol Beijing 100081 Peoples R China Hebei Univ Sch Cyber Secur & Comp Baoding 071000 Peoples R China

ISBN: (纸本)9783031189159;9783031189166

Recently, convolution neural network based methods have dominated the remote sensing image super-resolution (RSISR). However, most of them own complex network structures and a large number of network parameters, which is not friendly to computational resources limited scenarios. Besides, scale variations of objects in the remote sensing image are still challenging for most methods to generate high-quality super-resolution results. To this end, we propose a scale-aware group convolution (SGC) for RSISR. Specifically, each SGC firstly uses group convolutions with different dilation rates for extracting multi-scale features. Then, a scale-aware feature guidance approach and enhancement approach are leveraged to enhance the representation ability of different scale features. Based on SGC, a cascaded scale-aware distillation network (CSDN) is designed, which is composed of multiple SGC based cascade scale-aware distillation blocks (CSDBs). The output of each CSDB will be fused via the backward feature fusion module for final image super-resolution reconstruction. Extensive experiments are performed on the commonly-used UC Merced dataset. Quantitative and qualitative experiment results demonstrate the effectiveness of our method.

关键词： remote sensing image super-resolution Lightweight neural network Multi-scale feature learning Feature distillation

来源：评论

学校读者我要写书评

暂无评论

Unsupervised Cross-Domain Network Based on Bi-classifier for PolSAR image Classification 6

Unsupervised Cross-Domain Network Based on Bi-classifier for...

引用

6th International conference on Artificial Intelligence and pattern recognition, AIPR 2023

作者： Hua, Wenqiang Liu, Lin Sun, Nan Wang, Chen Xi'an University of Posts and Telecommunications China

ISBN: (纸本)9798400707674

Polarimetric Synthetic Aperture Radar (PolSAR) image classification is an important research content in the field of remote sensing. However, the lack of effective labeled data and the serious class distribution shift problem between different datasets hinder the application of deep learning networks in PolSAR image classification. Therefore, this paper proposes a Bi-classifier adversarial network based on covariance difference matrix and Convolutional Block Attention Module (CBAM) to solve the cross-domain problem in PolSAR images. Firstly, since PolSAR image processing generally uses the covariance matrix or coherence matrix of a single target pixel to construct the data set, which cannot effectively use the local information of PolSAR data,Therefore, the PolSAR coherence difference matrix P is proposed as the description of the PolSAR image difference information. Secondly, due to the different importance of features between neighboring pixels, the CBAM mechanism is introduced in order to exploit the spatial information of PolSAR data. Finally, in order to solve the cross-domain problem of PolSAR images more effectively, a bi-classifier adversarial network based on covariance difference matrix and CBAM is proposed. In this paper, we conduct experiments on public datasets, using images of San FranciscoII acquired in 2008 as the source domain, and images of San Francisco II acquired in 1992 as the target domain data. The experimental results show that the proposed method can effectively solve the cross-domain problem of PolSAR images. © 2023 ACM.

关键词： image classification

来源：评论

学校读者我要写书评

暂无评论

A Dual Attention Fusion Network for SAR-Optical Land Use Classification Based on Semantic Balance 7

A Dual Attention Fusion Network for SAR-Optical Land Use Cla...

引用

7th International conference on pattern recognition and Artificial Intelligence, PRAI 2024

作者： Yu, Ke Wang, Feng Shanghai200433 China Yiwu Research Institute Fudan University Yiwu322000 China

ISBN: (纸本)9798350350890

The rapid development of remote sensing technology has enhanced the acquisition of multi-modal data. Optical and synthetic aperture radar (SAR) images, as a uniquely complementary pair, have been proven highly effective in land use classification tasks, allowing for more accurate and detailed analysis of surface characteristics. However, existing approaches often rely on simple fusion methods and ignore the semantic relationships between modalities. Moreover, network models often focus on merging extensive contextual semantic information across channels to capture broader scene features, neglecting the connection between spatial textures and details. To address these issues, the dual attention optical and SAR image fusion network (DAOSNet) is proposed in this paper to balance semantic interactions between modalities and supplement local spatial information. DAOSN et consists of two key modules. Firstly, the dual attention fusion Transformer (DAFT) module can create cross-modal sensory fields to enhance global interactions between pixels in the spatial dimension. Additionally, the modal semantic balance (MSB) loss is introduced to promote deeper semantic collaboration between modalities and preserve the unique perspectives of each modality. The experimental results on the WHU-OPT -SAR dataset demonstrate that the proposed approach outperforms other fusion-based approaches, with an improvement of 1.4% in overall accuracy (OA), 10.4% in mean intersection over union (MIoU), and 1.8% in Kappa compared to the MCANet. © 2024 IEEE.

关键词： image fusion

来源：评论

学校读者我要写书评

暂无评论

A fuzzy spectral clustering algorithm for hyperspectral image classification

引用

IET image processing 2021年第12期15卷 2810-2817页

作者： Li, Kang Xu, Jindong Zhao, Tianyu Liu, Zhaowei YanTai Univ Sch Comp & Control Engn Yantai Peoples R China

Spectral clustering is an unsupervised clustering algorithm, and is widely used in the field of pattern recognition and computer vision due to its good clustering performance. However, the traditional spectral clustering algorithm is not suitable for large-scale data classification, such as hyperspectral remote sensing image, because of its high computational complexity, and it is difficult to characterize the inherent uncertainty of the hyperspectral remote sensing image. This paper uses fuzzy anchors to process hyperspectral image classification and proposes a novel spectral clustering algorithm based on fuzzy similarity measure. The proposed algorithm utilizes the fuzzy similarity measure to obtain the similarity between the data points and the anchors, and then gets the similarity matrix. Finally, spectral clustering is performed on the similarity matrix to compute the classification results. The experimental results on the hyperspectral remote sensing image data sets have demonstrated the effectiveness of the proposed algorithm, and the introduction of fuzzy similarity measure gives rise to a more robust similarity matrix. Compared with existing methods, the proposed algorithm has a better classification result on the hyperspectral remote sensing image, and the kappa coefficient obtained by the proposed algorithm is 2% higher than the traditional algorithms.

关键词： Instrumentation and techniques for geophysical, hydrospheric and lower atmosphere research Optical, image and video signal processing Computer vision and image processing techniques Data handling techniques Combinatorial mathematics

来源：评论

学校读者我要写书评

暂无评论

3rd World conference on Intelligent and 3D Technologies, WCI3DT 2024

3rd World Conference on Intelligent and 3D Technologies, WCI...

引用

3rd World conference on Intelligent and 3D Technologies, WCI3DT 2024

ISBN: (纸本)9789819791279

The proceedings contain 45 papers. The special focus in this conference is on Intelligent and 3D Technologies. The topics include: Multi-scale Point Cloud Shape Completion Network Based on Deep Learning;research on Application and Effect Evaluation of Product Innovation Management Based on Deep Learning;study on Deep Learning-Based Personalised Product Recommendation Model for Autonomous Question-and-Answer Robot;application of image Watermarking Technology Based on Deep Learning in Copyright Protection;research on the Combination of Building Structural Health Monitoring and Deep Learning image processing;electrical Equipment Prediction in a Variable Electromagnetic Field Using Deep Learning;state Monitoring and Fault Prediction of Wind Farm Transmission and Transformation Equipment Based on Deep Learning;application of Deep Learning Algorithms in the Innovation Ecosystem of Electric Power;optimization Strategy for Inventory Management Based on Machine Learning;intelligent Design and Evaluation of Aging Adaptable Public Spaces Based on Deep Learning;performance Optimization and Acceleration of Machine Learning Algorithms in Task Allocation of Mine Maintenance Robots;Application of Deep Learning to Improve the Performance of Automotive Electronic Control Unit (ECU);deep Learning-Based Scene Classification for remote sensing images;Improved Swarm Intelligence Optimization Algorithm Based on SL-Relu Activation Function Improvement Strategy and Its Application in Price Forecasting;Automatic Segmentation of Traumatic Penumbra in Rat Brain Based on Improved UNet++;A Brain Tumor Classification Method Based on ResNeXt-SESA Network;A LSTM Algorithm for Coastal City Cultural Scene Value Sustainable Development Forecast Improvement;pattern recognition in Archive Analysis Using Data Mining;building Crack Detection Method Based on Convolutional Neural Network.

关键词：

来源：评论

学校读者我要写书评

暂无评论

SAR Target Zero-Shot recognition with Optical image Assistance 2

SAR Target Zero-Shot Recognition with Optical Image Assistan...

引用

2nd IEEE International conference on Signal, Information and Data processing, ICSIDP 2024

作者： Chen, Jian Yong, Qifeng Du, Lan Zhou, Yashi Liu, Jie Xidian University National Key Laboratory of Radar Signal Processing Xi'an China Institute of Remote Sensing Satellite China Academy of Space Technology Beijing China

ISBN: (纸本)9798331515669

Most synthetic aperture radar (SAR) target recognition (ATR) methods focus on closed set conditions, which assumes that the testing set shares the same label space with the training set. However, new targets inevitably appear in practical testing phase. Closed-set SAR target recognition methods will misclassify those unseen new targets as the known targets in training phase, causing decision errors. Zero-shot recognition (ZSR) is an effective way to make predictions for unknown classes. Existing ZSR methods, which are primarily designed for optical images, typically construct an attribute knowledge base shared among known and unknown targets, and then learn the feature-attribute mapping using only known class data, which is subsequently applied to predict attributes of unknowns during testing and achieve class judgment. Unfortunately, the optical ZSR methods do not perform well in SAR data, due to the huge difficulty in the learning of feature-attribute mapping caused by the less accurate target representation of SAR images than the optical images. To overcome this problem, this paper proposes a ZSR method of SAR target with optical image assistance. In detail, on the basis of attribute knowledge base construction, the proposed method maps the known optical and SAR training data to the feature space through their respective feature extraction network, and uses feature aggregation loss to ensure that optical and SAR image features from the same category are close to each other, while those from different categories remain far apart. Based on the alignment of optical and SAR features, the feature-attribute mapping is learned using both SAR and optical data, leveraging the advantages of optical images in the easier and more accurate learning of feature-attribute mapping to promote the learning of SAR images. In testing phase, the learned SAR feature extraction network and feature-attribute mapping are used to predict unknown targets' attributes. Ultimately, the unknown

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

Lightweight Synthetic Aperture Radar image Transmission Facility Detection Method Based on YOLOv5 5

Lightweight Synthetic Aperture Radar Image Transmission Faci...

引用

5th International conference on Computer Vision, image and Deep Learning, CVIDL 2024

作者： Wen, Gang Zhou, Fangrong Yu, Hui Ma, Yi Wang, Guofang Wang, Yifan Geng, Hao Cao, Jun Kunming650217 China

ISBN: (纸本)9798350373820

Transmission facilities play a crucial role in ensuring electrical safety. Implementing remote sensing monitoring for transmission facilities in challenging terrains offers robust support for daily maintenance inspections and early accident warnings. In view of the small scale of power facility targets relative to surrounding terrain features in high-resolution Synthetic Aperture Radar (SAR) images, this paper proposes a lightweight transmission facility target recognition method based on an improved YOLOv5 network. The paper mainly focuses on network improvement and lightweight processing. In the YOLOv5 network, fused dilated convolutions and Convolutional Block Attention Module (CBAM) attention mechanisms are introduced to adjust the weights of small target receptive fields, enhancing the feature extraction of small target power facilities. The Filter Pruning via Geometric Median (FPGM) pruning is used to adjust the network architecture, remove redundant feature modules, and achieve lightweight models. The Varifocal loss is used as the loss function to balance positive and negative classes, highlight the contribution of positive examples, and improve model accuracy. The results show that after optimization, the accuracy can be improved to a maximum of 9 8. 2%. After pruning, the model size significantly decreases, and the inference time can be reduced to below 3 ms. Compared with current mainstream algorithms, the performance is greatly improved, which can meet the needs of transmission facility target detection in remote sensing SAR images. © 2024 IEEE.

关键词： Synthetic aperture radar

来源：评论

学校读者我要写书评

暂无评论

3rd World conference on Intelligent and 3D Technologies, WCI3DT 2024

3rd World Conference on Intelligent and 3D Technologies, WCI...

引用

3rd World conference on Intelligent and 3D Technologies, WCI3DT 2024

ISBN: (纸本)9789819791231

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on remote sensing image processing Model Based on Convolutional Neural Net 23

Research on Remote Sensing Image Processing Model Based on C...

引用

Proceedings of the 6th International conference on Information Technologies and Electrical Engineering

作者： Zhe Wang School of Information Engineering Wuhan University of Technology China

ISBN: (纸本)9798400708299

With the ongoing advancements in artificial intelligence technology, Convolutional Neural Networks (CNNs) have found widespread application in the realm of remote sensing image processing. They prove instrumental in tasks such as feature categorization, target detection, and change detection. Despite these strides, the continuous evolution of deep learning-based technology for remote sensing image processing confronts intricate and specific challenges, including the scarcity of high-quality remote sensing image data, strategies for handling multimodal data, and the detection of small targets in images. This paper proposes a remote sensing image processing model grounded in Convolutional Neural Networks to enhance the classification accuracy of remote sensing images. The approach incorporates the use of bilateral filtering to address interference factors in the image, preserving crucial image details. Furthermore, point operation-based histogram equalization is employed for image processing, facilitating the extraction of inconspicuous features within the image. The proposed model undergoes testing, and after more than 40 training iterations, it achieves a classification accuracy of over 90% for remote sensing images. The image processing results demonstrate effectiveness, showcasing high model recognition accuracy.

关键词： Convolutional neural network Deep learning remote sensing images Target recognition

来源：评论

学校读者我要写书评

暂无评论

MSDGAN: Multi-Scale Dilated Generative Adversarial Network for Smoke Removal and image Restoration 24

MSDGAN: Multi-Scale Dilated Generative Adversarial Network f...

引用

2024 5th International Artificial Intelligence and Blockchain conference, AIBC 2024

作者： Gwak, KyungMIn Rho, Young. J Tech University of Korea Siheung Korea Republic of

ISBN: (纸本)9798400710780

This paper presents a novel approach for smoke removal and image restoration using a Multi-Scale Dilated Generative Adversarial Network (MSDGAN). The presence of smoke in images poses significant challenges to both human perception and computer vision tasks, reducing visibility and complicating image processing tasks such as object recognition and segmentation. Traditional techniques like dehazing and defogging often fail to effectively address the irregular and transparent nature of smoke. To overcome these limitations, we propose a network architecture that integrates multi-scale dilated convolutions within a GAN framework, enabling the model to capture contextual information across various scales and improve the restoration of images affected by smoke. The MSDGAN architecture includes an encoder-decoder structure with skip connections, and it leverages perceptual loss and SSIM loss to maintain feature fidelity and reconstruct high-quality images. Extensive experiments demonstrate the effectiveness of the proposed network in removing smoke and restoring image quality across various challenging environments. The results show that MSDGAN outperforms existing dehazing and image restoration methods, offering superior performance as measured by L1, L2, SSIM, PSNR, and LPIPS metrics. This work provides a robust solution for enhancing image clarity in the presence of smoke, with potential applications in fields such as autonomous driving, surveillance, and remote sensing. Copyright © 2024 held by the owner/author(s). Publication rights licensed to ACM.

关键词： image reconstruction

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：