"Computer vision can make important contributions to the analysis of remote sensing satellite or aerial imagery. However, the resolution of early satellite imagery was not sufficient to provide useful spatial features. The situation is changing with the advent of very-high-spatial-resolution (VHR) imaging sensors. This change makes it possible to use computer vision techniques to perform analysis of man-made structures. Meanwhile, the development of multi-view imaging techniques allows the generation of accurate point clouds as ancillary knowledge. This dissertation aims at developing computer vision and machine learning algorithms for high resolution aerial imagery analysis in the context of application problems including debris detection, building detection and roof condition assessment. High resolution aerial imagery and point clouds were provided by Pictometry International for this study. Debris detection after natural disasters such as tornadoes, hurricanes or tsunamis, is needed for effective debris removal and allocation of limited resources. Significant advances in aerial image acquisition have greatly enabled the possibilities for rapid and automated detection of debris. In this dissertation, a robust debris detection algorithm is proposed. Large scale aerial images are partitioned into homogeneous regions by interactive segmentation. Debris areas are identified based on extracted texture features. Robust building detection is another important part of high resolution aerial imagery understanding. This dissertation develops a 3D scene classification algorithm for building detection using point clouds derived from multi-view imagery. Point clouds are divided into point clusters using Euclidean clustering. Individual point clusters are identified based on extracted spectral and 3D structural features. The inspection of roof condition is an important step in damage claim processing in the insurance industry. Automated roof condition assessment from remotely sensed images is proposed in this dissertation. Initially, texture classification and a bag-of-words model were applied to assess the roof condition using features derived from the whole rooftop. However, considering the complexity of residential rooftop, a more sophisticated method is proposed to divide the task into two stages: 1) roof segmentation, followed by 2) classification of segmented roof regions. Deep learning techniques are investigated for both segmentation and classification. A deep learned feature is proposed and applied in a region merging segmentation algorithm. A fine-tuned deep network is adopted for roof segment classification and found to achieve higher accuracy than traditional methods using hand-crafted features. Contributions of this study include the development of algorithms for debris detection using 2D images and building detection using 3D point clouds. For roof condition assessment, the solutions to this problem are explored in two directions: features derived from the whole rooftop and features extracted from each roof segments. Through our research, roof segmentation followed by segments classification was found to be a more promising method and the workflow processing developed and tested. Deep learning techniques are also investigated for both roof segmentation and segments classification. More unsupervised feature extraction techniques using deep learning can be explored in future work."--Abstract.

This book presents a comprehensive review of image processing methods, for the analysis of land use in residential areas. Combining a theoretical framework with highly practical applications, the book describes a system for the effective detection of single houses and streets in very high resolution. Topics and features: with a Foreword by Prof. Dr. Peter Reinartz of the German Aerospace Center; provides end-of-chapter summaries and review questions; presents a detailed review on remote sensing satellites; examines the multispectral information that can be obtained from satellite images, with a focus on vegetation and shadow-water indices; investigates methods for land-use classification, introducing precise graph theoretical measures over panchromatic images; addresses the problem of detecting residential regions; describes a house and street network-detection subsystem; concludes with a summary of the key ideas covered in the book.

Advancements in digital sensor technology, digital image analysis techniques, as well as computer software and hardware have brought together the fields of computer vision and photogrammetry, which are now converging towards sharing, to a great extent, objectives and algorithms. The potential for mutual benefits by the close collaboration and interaction of these two disciplines is great, as photogrammetric know-how can be aided by the most recent image analysis developments in computer vision, while modern quantitative photogrammetric approaches can support computer vision activities. Devising methodologies for automating the extraction of man-made objects (e.g. buildings, roads) from digital aerial or satellite imagery is an application where this cooperation and mutual support is already reaping benefits. The valuable spatial information collected using these interdisciplinary techniques is of improved qualitative and quantitative accuracy. This book offers a comprehensive selection of high-quality and in-depth contributions from world-wide leading research institutions, treating theoretical as well as implementational issues, and representing the state-of-the-art on this subject among the photogrammetric and computer vision communities.

Since 1994, the European Commission has undertaken various actions to expand the use of Earth observation (EO) from space in the Union and to stimulate value-added services based on the use of Earth observation satellite data.' By supporting research and technological development activities in this area, DG XII responded to the need to increase the cost-effectiveness of space derived environmental information. At the same time, it has contributed to a better exploitation of this unique technology, which is a key source of data for environmental monitoring from local to global scale. MAVIRIC is part of the investment made in the context of the Environ ment and Climate Programme (1994-1998) to strengthen applied techniques, based on a better understanding of the link between the remote sensing signal and the underlying bio- geo-physical processes. Translation of this scientific know-how into practical algorithms or methods is a priority in order to con vert more quickly, effectively and accurately space signals into geographical information. Now the availability of high spatial resolution satellite data is rapidly evolving and the fusion of data from different sensors including radar sensors is progressing well, the question arises whether existing machine vision approaches could be advantageously used by the remote sensing community. Automatic feature/object extraction from remotely sensed images looks very attractive in terms of processing time, standardisation and implementation of operational processing chains, but it remains highly complex when applied to natural scenes.

Advances in Aerial Sensing and Imaging This groundbreaking book is a comprehensive guide to the technology found in the complex field of aerial sensing and imaging, and the real-world challenges that stem from its growing significance and demand. The advent of unmanned aerial vehicles (UAVs), or drones, along with advancements in sensor technology and image processing techniques, has further enhanced the capabilities and applications of aerial sensing and imaging. These developments have opened up new research, innovation, and exploration avenues. Aerial sensing and imaging have rapidly evolved over the past few decades and have revolutionized several fields, including land cover and usage prediction, crop and livestock management, road accident monitoring, poverty estimation, defense, agriculture, forest fire detection, UAV security issues, and open parking management. This book provides a comprehensive understanding and knowledge of the underlying technology and its practical applications in different domains. Audience Computer science and artificial intelligence researchers working in the fields of aerial sensing and imaging, as well as professionals working in industries such as agriculture, geology, surveying, urban planning, disaster response, etc; this book provides them with practical guidance and instruction on how to apply aerial sensing and imaging for various purposes and stay up-to-date with the latest developments in the domain.

Satellite imagery analysis using deep learning methods, specifically convolutional neural networks (CNNs), has grown in popularity since 2012, with uses extending into the estimation of population, wealth, poverty, conflict, migration, education, and infrastructure, among other applications. This dissertation contributes to this body of literature in three parts. First, I explore the use of deep learning to overcome the sparsity, or complete lack, of accurate information regarding existing road infrastructure across much of the world. Using a novel labeled dataset generated by a custom-coded Android application, I show that a transfer learning approach can estimate road quality based on high-resolution satellite imagery with an accuracy of up to 80%. In the second chapter, I illustrate the vulnerability of this and related models to cyber intrusions (data poisoning), and propose a new technique to mitigate these vulnerabilities. The third chapter applies the lessons learned to propose a novel model architecture for spatiotemporal monitoring of industrial sites in inaccessible regions around the world, integrating high-resolution satellite imagery, a segmentation algorithm, and a pretrained deep learning framework to automatically detect and monitor individual industrial sites within the People's Republic of China. These three chapters advance our understanding of many of the challenges unique to computer vision in the context of satellite data, and provide some guidance on fruitful future directions.

With the recent advances in remote sensing technologies for Earth observation, many different remote sensors are collecting data with distinctive properties. The obtained data are so large and complex that analyzing them manually becomes impractical or even impossible. Therefore, understanding remote sensing images effectively, in connection with physics, has been the primary concern of the remote sensing research community in recent years. For this purpose, machine learning is thought to be a promising technique because it can make the system learn to improve itself. With this distinctive characteristic, the algorithms will be more adaptive, automatic, and intelligent. This book introduces some of the most challenging issues of machine learning in the field of remote sensing, and the latest advanced technologies developed for different applications. It integrates with multi-source/multi-temporal/multi-scale data, and mainly focuses on learning to understand remote sensing images. Particularly, it presents many more effective techniques based on the popular concepts of deep learning and big data to reach new heights of data understanding. Through reporting recent advances in the machine learning approaches towards analyzing and understanding remote sensing images, this book can help readers become more familiar with knowledge frontier and foster an increased interest in this field.