检索结果-内蒙古大学图书馆

Multi-Scale Mahalanobis Kernel-Based Support Vector Machine for Classification of High-Resolution remote sensing images

引用

COGNITIVE COMPUTATION 2021年第4期13卷 787-794页

作者： Sun, Genyun Rong, Xueqian Zhang, Aizhu Huang, Hui Rong, Jun Zhang, Xuming China Univ Petr East China Sch Geosci Qingdao 266580 Shandong Peoples R China Qingdao Natl Lab Marine Sci & Technol Lab Marine Mineral Resources Qingdao 266071 Peoples R China Key Lab Deep Oil & Gas Qingdao 266580 Peoples R China

Support vector machine (SVM) is a powerful cognitive and learning algorithm in the domain of pattern recognition and image classification. However, the generalization ability of SVM is limited when processing classification of high-resolution remote sensing images. One chief reason for this is that the Euclidean distance-based distance matrix in traditional SVM treats different samples equally and overlooks the global distribution of samples. To construct a more effective SVM-based classification method, this paper proposes a multi-scale Mahalanobis kernel-based SVM classifier. In this new method, we first introduce a Mahalanobis distance kernel to improve the global cognitive learning ability of SVM. Then, the Mahalanobis distance kernel is embedded to the multi-scale kernel learning (MSKL) to construct a novel multi-scale Mahalanobis kernel, in which the parameters are optimized by a bio-inspired algorithm, named differential evolution. Finally, the new method is extended to the classification of high-resolution remote sensing images based on the spatial-spectral features. The comparison experiments of five public UCI datasets and two high-resolution remote sensing images verify that the Mahalanobis distance-based method can obtain more accurate classification results than that of the Euclidean distance-based method. In addition, the proposed method produced the best classification results in all the experiments. The global cognitive learning ability of Mahalanobis distance-based method is stronger than that of the Euclidean distance-based method. In addition, this study indicates that the optimized MSKL are potential for the interpretation and understanding of complicated high-resolution remote sensing scene.

关键词： Support vector machine (SVM) Mahalanobis distance image classification High-resolution image Multi-scale kernel learning

来源：评论

学校读者我要写书评

暂无评论

Lightweight YOLOv4 Algorithm for remote sensing image Detection 14

Lightweight YOLOv4 Algorithm for Remote Sensing Image Detect...

引用

14th International Conference on Signal processing Systems, ICSPS 2022

作者： Ma, Li He, Tong Sun, Yong Hu, Bin-Bin College of Information Liaoning University Shenyang China

ISBN: (纸本)9798350336313

remote sensing images have the characteristics of complex backgrounds, high resolution, and small targets. Although the existing object detection algorithms can improve the detection accuracy, there are generally problems such as a large number of model parameters, high computational cost, and poor real-time performance. Aiming at the above problems, this paper designs a lightweight object detection algorithm GSC-YOLO based on YOLOv4 to achieve fast and accurate detection of remote sensing images. First, Ghostnet is used as the feature extraction network of GSC-YOLO to reduce the number of parameters and improve the detection speed;Secondly, the improved shuffle attention mechanism is introduced in the prediction head to make the model pay attention to important information and improve the detection accuracy;Finally, the Confidence Propagation Cluster algorithm CP-Cluster is used to post-process the prediction frame to improve the object recognition. Taking the preprocessed DOTA dataset as the experimental object, the experimental results show that the GSC-YOLO algorithm has a detection accuracy of 93.44%, a detection speed of 58 frames per second, and a model size of 43.65MB. Compared with the remote sensing image object detection algorithm based on YOLOv4, the detection accuracy is increased by 3.93%, the detection speed is increased by 1.87 times, and the model size is reduced by 5.62 times, which is more suitable for deployment on devices with limited resources. © 2022 IEEE.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Scene recognition of remote sensing image with residuals

Scene recognition of remote sensing image with residuals

引用

Consumer Life Tech (ICLT), IEEE

作者： Zhuyu Shao Shuguang Zhang Zhen Wang Xiaodong Chen Peng Li Heze Power Supply Company State Grid Shandong Electric Power Company Heze China

ISBN: (数字)9798331519339

ISBN: (纸本)9798331519346

Scene recognition of remote sensing images can be used for weapon guidance, forest fire prevention, etc., and has wide applications in environmental testing, urban planning and other fields. The development of deep learning technology has provided new possibilities for improving the accuracy of scene recognition in remote sensing images. Three network models for scene recognition in remote sensing images are designed based on residual networks and convolutional neural networks (CNNs). Taking the UCMerced_LandUse dataset as an example, the network depth is determined through pre-training. In the UCMerced_LandUse test set, the accuracy of the residual network model is 85% under 100 training cycles, which is about 50% higher than the residual network without BN layer and about 20% higher than the CNN model. By increasing the training cycle, the accuracy of the residual network model on all datasets reaches 96%. The residual network model designed in this paper does not significantly improve the accuracy in remote sensing image scene recognition tasks, but the model has the advantage of being lightweight. The depth of only 8 layers of residual blocks can be comparable to the accuracy of typical networks such as ResNet50, VGG16, MobileNet etc. The application prospects of lightweight deep learning technology in the field of remote sensing image processing are very broad, providing strong support for the further development of this field.

关键词： Training Deep learning image recognition Accuracy Translation Weapons Urban planning Convolutional neural networks remote sensing Residual neural networks

来源：评论

学校读者我要写书评

暂无评论

Deep semantic-aware remote sensing image deblurring

引用

SIGNAL processing 2023年 211卷

作者： Song, Zhenbo Zhang, Zhenyuan Fang, Feiyi Fan, Zhaoxin Lu, Jianfeng Nanjing Univ Sci & Technol Sch Comp Sci & Engn Nanjing 210094 Peoples R China Renmin Univ China Sch Informat Beijing 100872 Peoples R China

This paper addresses the problem of blind deblurring of single remote sensing (RS) images with deep neural networks. Most existing deep learning-based methods are migrated from natural image deblurring models, disregarding the domain gap to remote sensing images. Besides, the image deblurring problem is typically considered as an independent low-level image pre-processing, taking no account of downstream tasks, such as classification and segmentation. In this paper, we first present a novel decoder with a par-allel fusion stream for fusing multi-scale RS features and expanding the receptive field. Then, to generate high-quality sharp RS images, we propose to calculate the perceptual loss on an RS-pre-trained network instead of computing on VGG19 pre-trained on natural images i.e. imageNet. For bridging the RS im-age deblurring results to the downstream recognition tasks, we further propose a semantic loss, which is calculated on the last-layer feature map of an RS segmentation network. With extensive experiments con-ducted on public RS image datasets, we demonstrate that the proposed method improves results for RS image deblurring and achieves competitive performance both qualitatively and quantitatively. Moreover, downstream recognition experiments validate the superior quality of the recovered images over existing methods.& COPY;2023 Elsevier B.V. All rights reserved.

关键词： remote sensing image deblurring Deep learning Semantic supervision

来源：评论

学校读者我要写书评

暂无评论

Segmentation of Natural and Anthropogenic Objects by Panchromatic Satellite images Using Statistical Textural Features

引用

OPTOELECTRONICS INSTRUMENTATION AND DATA processing 2022年第2期58卷 167-179页

作者： Dmitriev, E., V Kondranin, T., V Zotov, S. A. Russian Acad Sci Marchuk Inst Numer Math Moscow 119333 Russia Moscow Inst Phys & Technol Dolgoprudnyi 141701 Moscow Oblast Russia

The problem of segmentation of natural and technogenic objects by panchromatic satellite images of ultra-high spatial resolution (95%) the main types of natural and anthropogenic objects, as well as various structures... 详细信息

The problem of segmentation of natural and technogenic objects by panchromatic satellite images of ultra-high spatial resolution (<1 m) using texture analysis is considered. The effectiveness of various statistical methods for extracting textural features is analyzed. On the basis of numerical experiments, methods have been identified that make it possible to segment with high accuracy (>95%) the main types of natural and anthropogenic objects, as well as various structures of the plant canopy. The TTSPCA method is proposed, which makes it possible to share the most informative features extracted using various statistical methods. The results of test calculations show that this method has higher accuracy (>99%) in solving the problem of texture segmentation compared to the considered standard approaches.

关键词： remote sensing pattern recognition texture features thematic processing classification segmentation

来源：评论

学校读者我要写书评

暂无评论

AUTOMATIC ENRICHMENT OF INDOOR 3D MODELS USING A DEEP LEARNING APPROACH BASED ON SINGLE imageS WITH UNKNOWN CAMERA POSES

AUTOMATIC ENRICHMENT OF INDOOR 3D MODELS USING A DEEP LEARNI...

引用

ISPRS WG V/7 Conference on Measurement, Visualisation and processing in BIM for Design and Construction Management II

作者： Jarzabek-Rychard, M. Maas, H-G Wroclaw Univ Environm & Life Sci Inst Geodesy & Geoinformat Wroclaw Poland Tech Univ Dresden Inst Photogrammetry & Remote Sensing Dresden Germany

3D building modeling is a diverse field of research with a multitude of challenges, where data integration is an inherent component. The intensively growing market of BIM-related consumer applications requires methods and algorithms that enable efficient updates of existing 3D models without the need for cost-intensive data capturing and repetitive reconstruction processes. We propose a novel approach for semantic enrichment of existing indoor models by window objects, based on amateur camera RGB images with unknown exterior orientation parameters. The core idea of the approach is the parallel estimation of image camera poses with semantic recognition of target objects and their automatic mapping onto a 3D vector model. The presented solution goes beyond pure texture matching and links deep learning detection techniques with camera pose estimation and 3D reconstruction. To evaluate the performance of our procedure, we compare the estimated camera parameters with reference data, obtaining median values of 13.8 cm for the camera position and 1.1 degrees for its orientation. Furthermore, a quality of 3D mapping is assessed based on the comparison to the reference 3D point cloud. All the windows presented in the data source were detected successfully, with a mean distance between both point sets equal to 3.6 cm. The experimental results prove that the presented approach achieves accurate integration of objects extracted from single images with an input 3D model, allowing for an effective increase of its semantic coverage.

关键词： Building Information Model (BIM) deep learning object recognition texture mapping camera pose estimation

来源：评论

学校读者我要写书评

暂无评论

Edge Segmentation based on Illumination Invariant Feature Detector Phase Congruency 5

Edge Segmentation based on Illumination Invariant Feature De...

引用

5th International Conference on Mobile Computing and Sustainable Informatics, ICMCSI 2024

作者： Patil, Rajendra V. Aggarwal, Renu Shivaji Deore, Shailesh Sunrise University Department of Computer Engineering Rajasthan Alwar India SunRise University Department of Computer Science and Engineering Rajasthan Alwar India SSVPS BSD College of Engineering Department of Computer Engineering Dhule India

ISBN: (纸本)9798350395235

image Segmentation is necessary stage in image investigation. In computer vision, image segmentation is the process of dividing digital image into multiple segments. Dividing the image into various parts simplifies the analysis of image. In numerous image analysis and computer visualization uses, including remote sensing, robot vision, medical image analysis, and pattern recognition, the subdivision of image into segments is an important preprocessing stage. Finding the outlines of objects in the images is the fundamental step in the segmentation of pictures. Zero crossover, Canny, Gradient dependent, zero cross, Laplacian are examples of traditional edge detection techniques. This study presents a novel method based on phase congruency for edge segmentation. As compared to traditional edge detection methods, advantage of phase congruency is its invariance to uneven illumination. To obtain edge segmentation outcomes, edge lines were obtained using phase congruency. Long edges give an approximate sense of the object's arrangement in the picture. Color likeness between edges is determined while short edges are removed. Color similar edges assigned same label. Performance analysis shows that proposed method produces good results for synthetic images and natural real world images in which objects are well separated by boundaries whereas it results in broken edges for images having complex structure. © 2024 IEEE.

关键词： image segmentation

来源：评论

学校读者我要写书评

暂无评论

Research on Target Detection Model of remote sensing image Based on Dual Attention Yolo

Research on Target Detection Model of Remote Sensing Image B...

引用

International Conference on Information Systems and Computer Aided Education (ICISCAE)

作者： Fan Xu College of Computer Science and Technology Ocean University of China Qingdao China

ISBN: (数字)9798350350760

ISBN: (纸本)9798350350777

With the widespread application of remote sensing technology, accurate and rapid detection and identification of targets in remote sensing images have become an important research area. Traditional target detection methods often face issues with low recognition accuracy and slow response when dealing with complex backgrounds or small targets. This study proposes an improved YOLO (You Only Look Once) model integrated with a dual attention mechanism to enhance the performance of target detection in remote sensing images. This mechanism utilizes spatial attention to increase the model's sensitivity to target locations, while channel attention enhances the recognition of target features, enabling more accurate identification of small and complex environment targets. The model was tested across multiple remote sensing image datasets, and the results show significant improvements in detection accuracy and real-time performance compared to traditional YOLO models. Additionally, the implementation of the model reveals the potential of the dual attention mechanism in improving the efficiency and accuracy of remote sensing image processing. With further optimization and adjustment, this model is expected to play a significant role in various application scenarios such as environmental monitoring, urban planning, and disaster assessment.

关键词： YOLO Accuracy Attention mechanisms Sensitivity Target recognition Face recognition Computational modeling Urban planning Real-time systems remote sensing

来源：评论

学校读者我要写书评

暂无评论

UAV/Drone-Based Identification of Geomorphic and Structural Anomalies as Probable Indicators of Economic Mineral Deposits in Girar Area of Lalitpur District, UP, (India)

引用

JOURNAL OF THE INDIAN SOCIETY OF remote sensing 2022年第5期50卷 937-947页

作者： Uniyal, Aniruddha Gond, Abhishek Tiwari, Tarun Pandit, Sanjay Gupta, Chirag Saikhom, Victor Dinkar, Gautam Remote Sensing Applicat Ctr UP Lucknow Uttar Pradesh India NESAC Umiam India DGM Lucknow Uttar Pradesh India

Satellite remote sensing data and UAV/Drone images have been conjunctively used for identification and mapping of probable indicators of economic mineral deposits in Girar area of Lalitpur district. Area of interest for UAV/Drone survey was selected from the digitally enhanced IRS P6 LISS-IV images. Very high-resolution (3.95 cm/pix) UAV/Drone images have been captured, and their photogrammetric processing and DEM generation were performed using Agisoft Metashape software. Subsequently, digital image enhancements such as principal component analysis and some filtering techniques viz. high-pass and low-pass filters using 3 x 3, 5 x 5 kernel have been applied on UAV images. Abrupt changes in regular linear and curvilinear pattern and also in the dimension of the landforms and the macro-level terrain information have also been detected from digitally enhanced and filtered drone data. Anomalous orientation, disposition and abrupt changes in the shape and pattern of landforms, particularly the ridges at some places, have been referred to as geomorphic anomalies. Lineaments with a few hundreds of meters of length have been delineated on spatially enhanced images and DEMs generated from UAV/Drone data. Sudden change in slope and relatively steeper valley walls are distinguishable on UAV data-based Digital Elevation Model (DEM), thereby vindicating the existence of lineaments. Intersections of ENE-WSW and NW-SE trending lineaments in Girar ridge area have been marked as structural anomalies. UAV/Drone images have not only yielded precise information about the spatial extent of anomalies (earlier been identified on IRS P6 LISS-IV satellite images), but have also helped in the identification of smaller geomorphic and structural anomalies, which were otherwise difficult to identify on satellite images. Further, the field investigations have revealed the rock outcrops of Banded Haematite Quartzite (BHQ) within these geomorphic and structural anomalies with BHQ considered to be

关键词： Unmanned aerial vehicle (UAV) image enhancement Lineament Mineral targeting

来源：评论

学校读者我要写书评

暂无评论

GLA-ViT: A Global Contextual Attention Based for remote sensing image Scene Classification

GLA-ViT: A Global Contextual Attention Based for Remote Sens...

引用

International Symposium on Parallel and Distributed processing with Applications, ISPA

作者： Hao Yuan Kun Liu Qinghao Peng Benli Zou Can Wang Meishu Li School of Information Science and Engineering University of Jinan Jinan China China High-Resolution Earth Observation System Shandong Center of Data and Application Jinan China

ISBN: (数字)9798331509712

ISBN: (纸本)9798331509729

remote sensing image Scene Classification (RSISC) is a critical area of research in remote sensing that is dedicated to categorizing remote sensing images into specific semantic classes based on their content. Despite the Convolutional Neural Networks’ (CNNs) effectiveness in feature extraction and classification, they face challenges in exploring contextual information necessary for a comprehensive understanding of remote sensing scenes. To address this limitation, a novel model based on the Transformer architecture has attracted considerable attention from researchers. The Transformer model has shown promise in extracting contextual information within remote sensing (RS) scenes. However, the original Transformer model’s performance in RS scene classification has been subpar due to the diverse content types and scales present in remote sensing scenes. To overcome these challenges, a context-based approach for remote sensing image scene classification called GLA-ViT is proposed. The GLA-ViT model integrates local and global self-attention modules to enhance its ability to capture intricate details and contextual information in remote sensing images, while also maintaining computational efficiency. Additionally, GLA-ViT incorporates an improved MFused-MBConv module to enhance the model’s recognition capabilities in complex scenes. Experimental results on well-known datasets such as AID, UCM, and NWPU-RESISC45 demonstrate the superior performance of GLA-ViT in remote sensing image scene classification related tasks.

关键词： Scene classification Computational modeling Face recognition Semantics Computer architecture Transformers Feature extraction Computational efficiency remote sensing Context modeling

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：