Recent advances in remote sensing image processing

Remote sensing image processing is nowadays a mature research area. The techniques developed in the field allow many real-life applications with great societal value. For instance, urban monitoring, fire detection or flood prediction can have a great impact on economical and environmental issues. To attain such objectives, the remote sensing community has turned into a multidisciplinary field of science that embraces physics, signal theory, computer science, electronics, and communications. From a machine learning and signal/image processing point of view, all the applications are tackled under specific formalisms, such as classification and clustering, regression and function approximation, image coding, restoration and enhancement, source unmixing, data fusion or feature selection and extraction. This paper serves as a survey of methods and applications, and reviews the last methodological advances in remote sensing image processing.

Kernel-based methods for change detection in remote sensing images

Nowadays, the joint exploitation of images acquired daily by remote sensing instruments and of images available from archives allows a detailed monitoring of the transitions occurring at the surface of the Earth. These modifications of the land cover generate spectral discrepancies that can be detected via the analysis of remote sensing images. Independently from the origin of the images and of type of surface change, a correct processing of such data implies the adoption of flexible, robust and possibly nonlinear method, to correctly account for the complex statistical relationships characterizing the pixels of the images. This Thesis deals with the development and the application of advanced statistical methods for multi-temporal optical remote sensing image processing tasks. Three different families of machine learning models have been explored and fundamental solutions for change detection problems are provided. In the first part, change detection with user supervision has been considered. In a first application, a nonlinear classifier has been applied with the intent of precisely delineating flooded regions from a pair of images. In a second case study, the spatial context of each pixel has been injected into another nonlinear classifier to obtain a precise mapping of new urban structures. In both cases, the user provides the classifier with examples of what he believes has changed or not. In the second part, a completely automatic and unsupervised method for precise binary detection of changes has been proposed. The technique allows a very accurate mapping without any user intervention, resulting particularly useful when readiness and reaction times of the system are a crucial constraint. In the third, the problem of statistical distributions shifting between acquisitions is studied. Two approaches to transform the couple of bi-temporal images and reduce their differences unrelated to changes in land cover are studied. The methods align the distributions of the images, so that the pixel-wise comparison could be carried out with higher accuracy. Furthermore, the second method can deal with images from different sensors, no matter the dimensionality of the data nor the spectral information content. This opens the doors to possible solutions for a crucial problem in the field: detecting changes when the images have been acquired by two different sensors.

Characterization of gravitational rock slope deformations at different spatial scales based on field, remote sensing and numerical approaches

Rock slope instabilities such as rock slides, rock avalanche or deep-seated gravitational slope deformations are widespread in Alpine valleys. These phenomena represent at the same time a main factor that control the mountain belts erosion and also a significant natural hazard that creates important losses to the mountain communities. However, the potential geometrical and dynamic connections linking outcrop and slope-scale instabilities are often unknown. A more detailed definition of the potential links will be essential to improve the comprehension of the destabilization processes and to dispose of a more complete hazard characterization of the rock instabilities at different spatial scales. In order to propose an integrated approach in the study of the rock slope instabilities, three main themes were analysed in this PhD thesis: (1) the inventory and the spatial distribution of rock slope deformations at regional scale and their influence on the landscape evolution, (2) the influence of brittle and ductile tectonic structures on rock slope instabilities development and (3) the characterization of hazard posed by potential rock slope instabilities through the development of conceptual instability models. To prose and integrated approach for the analyses of these topics, several techniques were adopted. In particular, high resolution digital elevation models revealed to be fundamental tools that were employed during the different stages of the rock slope instability assessment. A special attention was spent in the application of digital elevation model for detailed geometrical modelling of past and potential instabilities and for the rock slope monitoring at different spatial scales. Detailed field analyses and numerical models were performed to complete and verify the remote sensing approach. In the first part of this thesis, large slope instabilities in Rhone valley (Switzerland) were mapped in order to dispose of a first overview of tectonic and climatic factors influencing their distribution and their characteristics. Our analyses demonstrate the key influence of neotectonic activity and the glacial conditioning on the spatial distribution of the rock slope deformations. Besides, the volumes of rock instabilities identified along the main Rhone valley, were then used to propose the first estimate of the postglacial denudation and filling of the Rhone valley associated to large gravitational movements. In the second part of the thesis, detailed structural analyses of the Frank slide and the Sierre rock avalanche were performed to characterize the influence of brittle and ductile tectonic structures on the geometry and on the failure mechanism of large instabilities. Our observations indicated that the geometric characteristics and the variation of the rock mass quality associated to ductile tectonic structures, that are often ignored landslide study, represent important factors that can drastically influence the extension and the failure mechanism of rock slope instabilities. In the last part of the thesis, the failure mechanisms and the hazard associated to five potential instabilities were analysed in detail. These case studies clearly highlighted the importance to incorporate different analyses and monitoring techniques to dispose of reliable and hazard scenarios. This information associated to the development of a conceptual instability model represents the primary data for an integrated risk management of rock slope instabilities. - Les mouvements de versant tels que les chutes de blocs, les éboulements ou encore les phénomènes plus lents comme les déformations gravitaires profondes de versant représentent des manifestations courantes en régions montagneuses. Les mouvements de versant sont à la fois un des facteurs principaux contrôlant la destruction progressive des chaines orogéniques mais aussi un danger naturel concret qui peut provoquer des dommages importants. Pourtant, les phénomènes gravitaires sont rarement analysés dans leur globalité et les rapports géométriques et mécaniques qui lient les instabilités à l'échelle du versant aux instabilités locales restent encore mal définis. Une meilleure caractérisation de ces liens pourrait pourtant représenter un apport substantiel dans la compréhension des processus de déstabilisation des versants et améliorer la caractérisation des dangers gravitaires à toutes les échelles spatiales. Dans le but de proposer un approche plus globale à la problématique des mouvements gravitaires, ce travail de thèse propose trois axes de recherche principaux: (1) l'inventaire et l'analyse de la distribution spatiale des grandes instabilités rocheuses à l'échelle régionale, (2) l'analyse des structures tectoniques cassantes et ductiles en relation avec les mécanismes de rupture des grandes instabilités rocheuses et (3) la caractérisation des aléas rocheux par une approche multidisciplinaire visant à développer un modèle conceptuel de l'instabilité et une meilleure appréciation du danger . Pour analyser les différentes problématiques traitées dans cette thèse, différentes techniques ont été utilisées. En particulier, le modèle numérique de terrain s'est révélé être un outil indispensable pour la majorité des analyses effectuées, en partant de l'identification de l'instabilité jusqu'au suivi des mouvements. Les analyses de terrain et des modélisations numériques ont ensuite permis de compléter les informations issues du modèle numérique de terrain. Dans la première partie de cette thèse, les mouvements gravitaires rocheux dans la vallée du Rhône (Suisse) ont été cartographiés pour étudier leur répartition en fonction des variables géologiques et morphologiques régionales. En particulier, les analyses ont mis en évidence l'influence de l'activité néotectonique et des phases glaciaires sur la distribution des zones à forte densité d'instabilités rocheuses. Les volumes des instabilités rocheuses identifiées le long de la vallée principale ont été ensuite utilisés pour estimer le taux de dénudations postglaciaire et le remplissage de la vallée du Rhône lié aux grands mouvements gravitaires. Dans la deuxième partie, l'étude de l'agencement structural des avalanches rocheuses de Sierre (Suisse) et de Frank (Canada) a permis de mieux caractériser l'influence passive des structures tectoniques sur la géométrie des instabilités. En particulier, les structures issues d'une tectonique ductile, souvent ignorées dans l'étude des instabilités gravitaires, ont été identifiées comme des structures très importantes qui contrôlent les mécanismes de rupture des instabilités à différentes échelles. Dans la dernière partie de la thèse, cinq instabilités rocheuses différentes ont été étudiées par une approche multidisciplinaire visant à mieux caractériser l'aléa et à développer un modèle conceptuel trois dimensionnel de ces instabilités. A l'aide de ces analyses on a pu mettre en évidence la nécessité d'incorporer différentes techniques d'analyses et de surveillance pour une gestion plus objective du risque associée aux grandes instabilités rocheuses.

A Survey of Active Learning Algorithms for Supervised Remote Sensing Image Classification

Defining an efficient training set is one of the most delicate phases for the success of remote sensing image classification routines. The complexity of the problem, the limited temporal and financial resources, as well as the high intraclass variance can make an algorithm fail if it is trained with a suboptimal dataset. Active learning aims at building efficient training sets by iteratively improving the model performance through sampling. A user-defined heuristic ranks the unlabeled pixels according to a function of the uncertainty of their class membership and then the user is asked to provide labels for the most uncertain pixels. This paper reviews and tests the main families of active learning algorithms: committee, large margin, and posterior probability-based. For each of them, the most recent advances in the remote sensing community are discussed and some heuristics are detailed and tested. Several challenging remote sensing scenarios are considered, including very high spatial resolution and hyperspectral image classification. Finally, guidelines for choosing the good architecture are provided for new and/or unexperienced user.

Structured output SVM for remote sensing image classification

In the recent years, kernel methods have revealed very powerful tools in many application domains in general and in remote sensing image classification in particular. The special characteristics of remote sensing images (high dimension, few labeled samples and different noise sources) are efficiently dealt with kernel machines. In this paper, we propose the use of structured output learning to improve remote sensing image classification based on kernels. Structured output learning is concerned with the design of machine learning algorithms that not only implement input-output mapping, but also take into account the relations between output labels, thus generalizing unstructured kernel methods. We analyze the framework and introduce it to the remote sensing community. Output similarity is here encoded into SVM classifiers by modifying the model loss function and the kernel function either independently or jointly. Experiments on a very high resolution (VHR) image classification problem shows promising results and opens a wide field of research with structured output kernel methods.

Structural characterization of rockfall sources in Yosemite Valley from remote sensing data and field surveys

Yosemite Valley poses significant rockfall hazard and related risk due to its glacially steepened walls and approximately 4 million visitors annually. To assess rockfall hazard, it is necessary to evaluate the geologic structure that contributes to the destabilization of rockfall sources and locate the most probable future source areas. Coupling new remote sensing techniques (Terrestrial Laser Scanning, Aerial Laser Scanning) and traditional field surveys, we investigated the regional geologic and structural setting, the orientation of the primary discontinuity sets for large areas of Yosemite Valley, and the specific discontinuity sets present at active rockfall sources. This information, combined with better understanding of the geologic processes that contribute to the progressive destabilization and triggering of granitic rock slabs, contributes to a more accurate rockfall susceptibility assessment for Yosemite Valley and elsewhere.

Active learning methods for remote sensing image classification

In this paper, we propose two active learning algorithms for semiautomatic definition of training samples in remote sensing image classification. Based on predefined heuristics, the classifier ranks the unlabeled pixels and automatically chooses those that are considered the most valuable for its improvement. Once the pixels have been selected, the analyst labels them manually and the process is iterated. Starting with a small and nonoptimal training set, the model itself builds the optimal set of samples which minimizes the classification error. We have applied the proposed algorithms to a variety of remote sensing data, including very high resolution and hyperspectral images, using support vector machines. Experimental results confirm the consistency of the methods. The required number of training samples can be reduced to 10% using the methods proposed, reaching the same level of accuracy as larger data sets. A comparison with a state-of-the-art active learning method, margin sampling, is provided, highlighting advantages of the methods proposed. The effect of spatial resolution and separability of the classes on the quality of the selection of pixels is also discussed.

Semi-supervised remote sensing image classification with cluster kernels

A semisupervised support vector machine is presented for the classification of remote sensing images. The method exploits the wealth of unlabeled samples for regularizing the training kernel representation locally by means of cluster kernels. The method learns a suitable kernel directly from the image and thus avoids assuming a priori signal relations by using a predefined kernel structure. Good results are obtained in image classification examples when few labeled samples are available. The method scales almost linearly with the number of unlabeled samples and provides out-of-sample predictions.

Domain Adaptation in remote sensing: increasing the portability of land-cover classifiers

Among the types of remote sensing acquisitions, optical images are certainly one of the most widely relied upon data sources for Earth observation. They provide detailed measurements of the electromagnetic radiation reflected or emitted by each pixel in the scene. Through a process termed supervised land-cover classification, this allows to automatically yet accurately distinguish objects at the surface of our planet. In this respect, when producing a land-cover map of the surveyed area, the availability of training examples representative of each thematic class is crucial for the success of the classification procedure. However, in real applications, due to several constraints on the sample collection process, labeled pixels are usually scarce. When analyzing an image for which those key samples are unavailable, a viable solution consists in resorting to the ground truth data of other previously acquired images. This option is attractive but several factors such as atmospheric, ground and acquisition conditions can cause radiometric differences between the images, hindering therefore the transfer of knowledge from one image to another. The goal of this Thesis is to supply remote sensing image analysts with suitable processing techniques to ensure a robust portability of the classification models across different images. The ultimate purpose is to map the land-cover classes over large spatial and temporal extents with minimal ground information. To overcome, or simply quantify, the observed shifts in the statistical distribution of the spectra of the materials, we study four approaches issued from the field of machine learning. First, we propose a strategy to intelligently sample the image of interest to collect the labels only in correspondence of the most useful pixels. This iterative routine is based on a constant evaluation of the pertinence to the new image of the initial training data actually belonging to a different image. Second, an approach to reduce the radiometric differences among the images by projecting the respective pixels in a common new data space is presented. We analyze a kernel-based feature extraction framework suited for such problems, showing that, after this relative normalization, the cross-image generalization abilities of a classifier are highly increased. Third, we test a new data-driven measure of distance between probability distributions to assess the distortions caused by differences in the acquisition geometry affecting series of multi-angle images. Also, we gauge the portability of classification models through the sequences. In both exercises, the efficacy of classic physically- and statistically-based normalization methods is discussed. Finally, we explore a new family of approaches based on sparse representations of the samples to reciprocally convert the data space of two images. The projection function bridging the images allows a synthesis of new pixels with more similar characteristics ultimately facilitating the land-cover mapping across images.