Deep Convolutional Neural Networks for Object Extraction from High Spatial Resolution Remotely Sensed Imagery PDF Download
Are you looking for read ebook online? Search for your book and save it on your Kindle device, PC, phones or tablets. Download Deep Convolutional Neural Networks for Object Extraction from High Spatial Resolution Remotely Sensed Imagery PDF full book. Access full book title Deep Convolutional Neural Networks for Object Extraction from High Spatial Resolution Remotely Sensed Imagery by Yuanming Shu. Download full books in PDF and EPUB format.
Author: Yuanming Shu Publisher: ISBN: Category : Languages : en Pages : 119
Book Description
Developing methods to automatically extract objects from high spatial resolution (HSR) remotely sensed imagery on a large scale is crucial for supporting land user and land cover (LULC) mapping with HSR imagery. However, this task is notoriously challenging. Deep learning, a recent breakthrough in machine learning, has shed light on this problem. The goal of this thesis is to develop a deep insight into the use of deep learning to develop reliable automated object extraction methods for applications with HSR imagery. The thesis starts by re-examining the knowledge the remote sensing community has achieved on the problem, but in the context of deep learning. Attention is given to object-based image analysis (OBIA) methods, which are currently considered to be the prevailing framework for this problem and have had a far-reaching impact on the history of remote sensing. In contrast to common beliefs, experiments show that object-based methods suffer seriously from ill-defined image segmentation. They are less effective at leveraging the power of the features learned by deep convolutional neural networks (CNNs) than conventionally patch-based methods. This thesis then studies ways to further improve the accuracy of object extraction with deep CNNs. Given that vector maps are required as the final format in many applications, the focus is on addressing the issues of generating high-quality vector maps with deep CNNs. A method combining bottom-up deep CNN prediction with top-down object modeling is proposed for building extraction. This method also exhibits the potential to extend to other objects of interest. Experiments show that implementing the proposed method on a single GPU results in the capability of processing 756 km2 of 12 cm aerial images in about 30 hours. By post-editing on top of the resulting automated extraction, high-quality building vector maps can be produced about 4-times faster than conventional manual digitization methods.
Author: Yuanming Shu Publisher: ISBN: Category : Languages : en Pages : 119
Book Description
Developing methods to automatically extract objects from high spatial resolution (HSR) remotely sensed imagery on a large scale is crucial for supporting land user and land cover (LULC) mapping with HSR imagery. However, this task is notoriously challenging. Deep learning, a recent breakthrough in machine learning, has shed light on this problem. The goal of this thesis is to develop a deep insight into the use of deep learning to develop reliable automated object extraction methods for applications with HSR imagery. The thesis starts by re-examining the knowledge the remote sensing community has achieved on the problem, but in the context of deep learning. Attention is given to object-based image analysis (OBIA) methods, which are currently considered to be the prevailing framework for this problem and have had a far-reaching impact on the history of remote sensing. In contrast to common beliefs, experiments show that object-based methods suffer seriously from ill-defined image segmentation. They are less effective at leveraging the power of the features learned by deep convolutional neural networks (CNNs) than conventionally patch-based methods. This thesis then studies ways to further improve the accuracy of object extraction with deep CNNs. Given that vector maps are required as the final format in many applications, the focus is on addressing the issues of generating high-quality vector maps with deep CNNs. A method combining bottom-up deep CNN prediction with top-down object modeling is proposed for building extraction. This method also exhibits the potential to extend to other objects of interest. Experiments show that implementing the proposed method on a single GPU results in the capability of processing 756 km2 of 12 cm aerial images in about 30 hours. By post-editing on top of the resulting automated extraction, high-quality building vector maps can be produced about 4-times faster than conventional manual digitization methods.
Author: Florence Tupin Publisher: John Wiley & Sons ISBN: 1118898923 Category : Technology & Engineering Languages : en Pages : 277
Book Description
Dedicated to remote sensing images, from their acquisition to their use in various applications, this book covers the global lifecycle of images, including sensors and acquisition systems, applications such as movement monitoring or data assimilation, and image and data processing. It is organized in three main parts. The first part presents technological information about remote sensing (choice of satellite orbit and sensors) and elements of physics related to sensing (optics and microwave propagation). The second part presents image processing algorithms and their specificities for radar or optical, multi and hyper-spectral images. The final part is devoted to applications: change detection and analysis of time series, elevation measurement, displacement measurement and data assimilation. Offering a comprehensive survey of the domain of remote sensing imagery with a multi-disciplinary approach, this book is suitable for graduate students and engineers, with backgrounds either in computer science and applied math (signal and image processing) or geo-physics. About the Authors Florence Tupin is Professor at Telecom ParisTech, France. Her research interests include remote sensing imagery, image analysis and interpretation, three-dimensional reconstruction, and synthetic aperture radar, especially for urban remote sensing applications. Jordi Inglada works at the Centre National d’Études Spatiales (French Space Agency), Toulouse, France, in the field of remote sensing image processing at the CESBIO laboratory. He is in charge of the development of image processing algorithms for the operational exploitation of Earth observation images, mainly in the field of multi-temporal image analysis for land use and cover change. Jean-Marie Nicolas is Professor at Telecom ParisTech in the Signal and Imaging department. His research interests include the modeling and processing of synthetic aperture radar images.
Author: Yakoub Bazi Publisher: MDPI ISBN: 3036509860 Category : Science Languages : en Pages : 438
Book Description
The rapid growth of the world population has resulted in an exponential expansion of both urban and agricultural areas. Identifying and managing such earthly changes in an automatic way poses a worth-addressing challenge, in which remote sensing technology can have a fundamental role to answer—at least partially—such demands. The recent advent of cutting-edge processing facilities has fostered the adoption of deep learning architectures owing to their generalization capabilities. In this respect, it seems evident that the pace of deep learning in the remote sensing domain remains somewhat lagging behind that of its computer vision counterpart. This is due to the scarce availability of ground truth information in comparison with other computer vision domains. In this book, we aim at advancing the state of the art in linking deep learning methodologies with remote sensing image processing by collecting 20 contributions from different worldwide scientists and laboratories. The book presents a wide range of methodological advancements in the deep learning field that come with different applications in the remote sensing landscape such as wildfire and postdisaster damage detection, urban forest mapping, vine disease and pavement marking detection, desert road mapping, road and building outline extraction, vehicle and vessel detection, water identification, and text-to-image matching.
Author: Taskin Kavzoglu Publisher: MDPI ISBN: 3039438271 Category : Science Languages : en Pages : 256
Book Description
Artificial neural networks (ANNs) and evolutionary computation methods have been successfully applied in remote sensing applications since they offer unique advantages for the analysis of remotely-sensed images. ANNs are effective in finding underlying relationships and structures within multidimensional datasets. Thanks to new sensors, we have images with more spectral bands at higher spatial resolutions, which clearly recall big data problems. For this purpose, evolutionary algorithms become the best solution for analysis. This book includes eleven high-quality papers, selected after a careful reviewing process, addressing current remote sensing problems. In the chapters of the book, superstructural optimization was suggested for the optimal design of feedforward neural networks, CNN networks were deployed for a nanosatellite payload to select images eligible for transmission to ground, a new weight feature value convolutional neural network (WFCNN) was applied for fine remote sensing image segmentation and extracting improved land-use information, mask regional-convolutional neural networks (Mask R-CNN) was employed for extracting valley fill faces, state-of-the-art convolutional neural network (CNN)-based object detection models were applied to automatically detect airplanes and ships in VHR satellite images, a coarse-to-fine detection strategy was employed to detect ships at different sizes, and a deep quadruplet network (DQN) was proposed for hyperspectral image classification.
Author: Mengge Chen Publisher: ISBN: Category : Artificial intelligence Languages : en Pages : 94
Book Description
The past decades have witnessed a significant change in human societies with a fast pace and rapid urbanization. The boom of urbanization is contributed by the influx of people to the urban area and comes with building construction and deconstruction. The estimation of both residential and industrial buildings is important to reveal and demonstrate the human activities of the regions. As a result, it is essential to effectively and accurately detect the buildings in urban areas for urban planning and population monitoring. The automatic building detection method in remote sensing has always been a challenging task, because small targets cannot be identified in images with low resolution, as well as the complexity in the various scales, structure, and colours of urban buildings. However, the development of techniques improves the performance of the building detection task, by taking advantage of the accessibility of very high-resolution (VHR) remotely sensed images and the innovation of object detection methods. The purpose of this study is to develop a framework for the automatic detection of urban buildings from the VHR remotely sensed imagery at a large scale by using the state-of-art deep learning network. The thesis addresses the research gaps and difficulties as well as the achievements in building detection. The conventional hand-crafted methods, machine learning methods, and deep learning methods are reviewed and discussed. The proposed method employs a deep convolutional neural network (CNN) for building detection. Two input datasets with different spatial resolutions were used to train and validate the CNN model, and a testing dataset was used to evaluate the performance of the proposed building detection method. The experiment result indicates that the proposed method performs well at both building detection and outline segmentation task with a total precision of 0.92, a recall of 0.866, an F1-score of 0.891. In conclusion, this study proves the feasibility of CNN on solving building detection challenges using VHR remotely sensed imagery.
Author: Grégoire Montavon Publisher: Springer ISBN: 3642352898 Category : Computers Languages : en Pages : 753
Book Description
The twenty last years have been marked by an increase in available data and computing power. In parallel to this trend, the focus of neural network research and the practice of training neural networks has undergone a number of important changes, for example, use of deep learning machines. The second edition of the book augments the first edition with more tricks, which have resulted from 14 years of theory and experimentation by some of the world's most prominent neural network researchers. These tricks can make a substantial difference (in terms of speed, ease of implementation, and accuracy) when it comes to putting algorithms to work on real problems.
Author: Saurabh Prasad Publisher: Springer Nature ISBN: 3030386171 Category : Computers Languages : en Pages : 464
Book Description
This book reviews the state of the art in algorithmic approaches addressing the practical challenges that arise with hyperspectral image analysis tasks, with a focus on emerging trends in machine learning and image processing/understanding. It presents advances in deep learning, multiple instance learning, sparse representation based learning, low-dimensional manifold models, anomalous change detection, target recognition, sensor fusion and super-resolution for robust multispectral and hyperspectral image understanding. It presents research from leading international experts who have made foundational contributions in these areas. The book covers a diverse array of applications of multispectral/hyperspectral imagery in the context of these algorithms, including remote sensing, face recognition and biomedicine. This book would be particularly beneficial to graduate students and researchers who are taking advanced courses in (or are working in) the areas of image analysis, machine learning and remote sensing with multi-channel optical imagery. Researchers and professionals in academia and industry working in areas such as electrical engineering, civil and environmental engineering, geosciences and biomedical image processing, who work with multi-channel optical data will find this book useful.
Author: Mohammad Awrangjeb Publisher: MDPI ISBN: 3039283820 Category : Science Languages : en Pages : 442
Book Description
Building extraction from remote sensing data plays an important role in urban planning, disaster management, navigation, updating geographic databases, and several other geospatial applications. Even though significant research has been carried out for more than two decades, the success of automatic building extraction and modeling is still largely impeded by scene complexity, incomplete cue extraction, and sensor dependency of data. Most recently, deep neural networks (DNN) have been widely applied for high classification accuracy in various areas including land-cover and land-use classification. Therefore, intelligent and innovative algorithms are needed for the success of automatic building extraction and modeling. This Special Issue focuses on newly developed methods for classification and feature extraction from remote sensing data for automatic building extraction and 3D
Author: Mohammad Rezaee Publisher: ISBN: Category : Languages : en Pages : 0
Book Description
The incorporation of spatial and spectral information within multispectral satellite images is the key for accurate land cover mapping, specifically for discrimination of heterogeneous land covers. Traditional methods only use basic features, either spatial features (e.g. edges or gradients) or spectral features (e.g. mean value of Digital Numbers or Normalized Difference Vegetation Index (NDVI)) for land cover classification. These features are called low level features and are generated manually (through so-called feature engineering). Since feature engineering is manual, the design of proper features is time-consuming, only low-level features in the information hierarchy can usually be extracted, and the feature extraction is application-based (i.e., different applications need to extract different features). In contrast to traditional land-cover classification methods, Deep Learning (DL),adapting the artificial neural network (ANN) into a deep structure, can automatically generate the necessary high-level features for improving classification without being limited to low-level features. The higher-level features (e.g. complex shapes and textures) can be generated by combining low-level features through different level of processing. However, despite recent advances of DL for various computer vision tasks, especially for convolutional neural networks (CNNs) models, the potential of using DL for land-cover classification of multispectral remote sensing (RS) images have not yet been thoroughly explored. The main reason is that a DL network needs to be trained using a huge number of images from a large scale of datasets. Such training datasets are not usually available in RS. The only few available training datasets are either for object detection in an urban area, or for scene labeling. In addition, the available datasets are mostly used for land-cover classification based on spatial features. Therefore, the incorporation of the spectral and spatial features has not been studied comprehensively yet. This PhD research aims to mitigate challenges in using DL for RS land cover mapping/object detection by (1) decreasing the dependency of DL to the large training datasets, (2) adapting and improving the efficiency and accuracy of deep CNNs for heterogeneous classification, (3) incorporating all of the spectral bands in satellite multispectral images into the processing, and (4) designing a specific CNN network that can be used for a faster and more accurate detection of heterogeneous land covers with fewer amount of training datasets. The new developments are evaluated in two case studies, i.e. wetland detection and tree species detection, where high resolution multispectral satellite images are used. Such land-cover classifications are considered as challenging tasks in the literature. The results show that our new solution works reliably under a wide variety of conditions. Furthermore, we are releasing the two large-scale wetland and tree species detection datasets to the public in order to facilitate future research, and to compare with other methods.
Author: Xiaojun X. Yang Publisher: John Wiley & Sons ISBN: 111962584X Category : Technology & Engineering Languages : en Pages : 532
Book Description
Urban Remote Sensing The second edition of Urban Remote Sensing is a state-of-the-art review of the latest progress in the subject. The text examines how evolving innovations in remote sensing allow to deliver the critical information on cities in a timely and cost-effective way to support various urban management activities and the scientific research on urban morphology, socio-environmental dynamics, and sustainability. Chapters are written by leading scholars from a variety of disciplines including remote sensing, GIS, geography, urban planning, environmental science, and sustainability science, with case studies predominately drawn from North America and Europe. A review of the essential and emerging research areas in urban remote sensing including sensors, techniques, and applications, especially some critical issues that are shifting the directions in urban remote sensing research. Illustrated in full color throughout, including numerous relevant case studies and extensive discussions of important concepts and cutting-edge technologies to enable clearer understanding for non-technical audiences. Urban Remote Sensing, Second Edition will be of particular interest to upper-division undergraduate and graduate students, researchers and professionals working in the fields of remote sensing, geospatial information, and urban & environmental planning.
Author: Yuanming Shu
Author: Yuanming Shu
Author: Florence Tupin
Author: Yakoub Bazi
Author: Taskin Kavzoglu
Author: Mengge Chen
Author: Grégoire Montavon
Author: Saurabh Prasad
Author: Mohammad Awrangjeb
Author: Mohammad Rezaee
Author: Xiaojun X. Yang