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.

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.

Stability analysis of the 2007 Chehalis lake landslide based on long-range terrestrial photogrammetry and airborne LIDAR data

On December 4th 2007, a 3-Mm3 landslide occurred along the northwestern shore of Chehalis Lake. The initiation zone is located at the intersection of the main valley slope and the northern sidewall of a prominent gully. The slope failure caused a displacement wave that ran up to 38 m on the opposite shore of the lake. The landslide is temporally associated with a rain-on-snow meteorological event which is thought to have triggered it. This paper describes the Chehalis Lake landslide and presents a comparison of discontinuity orientation datasets obtained using three techniques: field measurements, terrestrial photogrammetric 3D models and an airborne LiDAR digital elevation model to describe the orientation and characteristics of the five discontinuity sets present. The discontinuity orientation data are used to perform kinematic, surface wedge limit equilibrium and three-dimensional distinct element analyses. The kinematic and surface wedge analyses suggest that the location of the slope failure (intersection of the valley slope and a gully wall) has facilitated the development of the unstable rock mass which initiated as a planar sliding failure. Results from the three-dimensional distinct element analyses suggest that the presence, orientation and high persistence of a discontinuity set dipping obliquely to the slope were critical to the development of the landslide and led to a failure mechanism dominated by planar sliding. The three-dimensional distinct element modelling also suggests that the presence of a steeply dipping discontinuity set striking perpendicular to the slope and associated with a fault exerted a significant control on the volume and extent of the failed rock mass but not on the overall stability of the slope.

Waveform inversion of crosshole georadar data: influence of source wavelet variability and the suitability of a single wavelet assumption

Waveform-based tomographic imaging of crosshole georadar data is a powerful method to investigate the shallow subsurface because of its ability to provide images of electrical properties in near-surface environments with unprecedented spatial resolution. A critical issue with waveform inversion is the a priori unknown source signal. Indeed, the estimation of the source pulse is notoriously difficult but essential for the effective application of this method. Here, we explore the viability and robustness of a recently proposed deconvolution-based procedure to estimate the source pulse during waveform inversion of crosshole georadar data, where changes in wavelet shape with location as a result of varying near-field conditions and differences in antenna coupling may be significant. Specifically, we examine whether a single, average estimated source current function can adequately represent the pulses radiated at all transmitter locations during a crosshole georadar survey, or whether a separate source wavelet estimation should be performed for each transmitter gather. Tests with synthetic and field data indicate that remarkably good tomographic reconstructions can be obtained using a single estimated source pulse when moderate to strong variability exists in the true source signal with antenna location. Only in the case of very strong variability in the true source pulse are tomographic reconstructions clearly improved by estimating a different source wavelet for each transmitter location.

Architectural and Functional Similarities between Trimeric ATP-Gated P2X Receptors and Acid-Sensing Ion Channels

ATP-gated P2X receptors and acid-sensing ion channels are two distinct ligand-gated ion channels that assemble into trimers. They are involved in many important physiological functions such as pain sensation and are recognized as important therapeutic targets. They have unrelated primary structures and respond to different ligands (ATP and protons) and are thus considered as two different ion channels. As a consequence, comparisons of the biophysical properties and underlying mechanisms have only been rarely made between these two channels. However, the recent determination of their molecular structures by X-ray crystallography has revealed unexpected parallels in the architecture of the two pores, providing a basis for possible functional analogies. In this review, we analyze the structural and functional similarities that are shared by these trimeric ion channels, and we outline key unanswered questions that, if addressed experimentally, may help us to elucidate how two unrelated ion channels have adopted a similar fold of the pore.

Statistical assessment of dataset shift and model portability in multi-angle in-track image acquisitions

In this study we propose an evaluation of the angular effects altering the spectral response of the land-cover over multi-angle remote sensing image acquisitions. The shift in the statistical distribution of the pixels observed in an in-track sequence of WorldView-2 images is analyzed by means of a kernel-based measure of distance between probability distributions. Afterwards, the portability of supervised classifiers across the sequence is investigated by looking at the evolution of the classification accuracy with respect to the changing observation angle. In this context, the efficiency of various physically and statistically based preprocessing methods in obtaining angle-invariant data spaces is compared and possible synergies are discussed.

Classification of very high spatial resolution imagery using mathematical morphology and support vector machines

We investigate the relevance of morphological operators for the classification of land use in urban scenes using submetric panchromatic imagery. A support vector machine is used for the classification. Six types of filters have been employed: opening and closing, opening and closing by reconstruction, and opening and closing top hat. The type and scale of the filters are discussed, and a feature selection algorithm called recursive feature elimination is applied to decrease the dimensionality of the input data. The analysis performed on two QuickBird panchromatic images showed that simple opening and closing operators are the most relevant for classification at such a high spatial resolution. Moreover, mixed sets combining simple and reconstruction filters provided the best performance. Tests performed on both images, having areas characterized by different architectural styles, yielded similar results for both feature selection and classification accuracy, suggesting the generalization of the feature sets highlighted.

Characterization and management of rockslide hazard at Turtle Mountain, Alberta, Canada

In 1903, more than 30 million m3 of rock fell from the east slopes of Turtle Mountain in Alberta, Canada, causing a rock avalanche that killed about 70 people in the town of Frank. The Alberta Government, in response to continuing instabilities at the crest of the mountain, established a sophisticated field laboratory where state-of-the-art monitoring techniques have been installed and tested as part of an early-warning system. In this chapter, we provide an overview of the causes, trigger, and extreme mobility of the landslide. We then present new data relevant to the characterization and detection of the present-day instabilities on Turtle Mountain. Fourteen potential instabilities have been identified through field mapping and remote sensing. Lastly, we provide a detailed review of the different in-situ and remote monitoring systems that have been installed on the mountain. The implications of the new data for the future stability of Turtle Mountain and related landslide runout, and for monitoring strategies and risk management, are discussed.

Urban image classification with semisupervised multiscale cluster kernels

This paper presents a semisupervised support vector machine (SVM) that integrates the information of both labeled and unlabeled pixels efficiently. Method's performance is illustrated in the relevant problem of very high resolution image classification of urban areas. The SVM is trained with the linear combination of two kernels: a base kernel working only with labeled examples is deformed by a likelihood kernel encoding similarities between labeled and unlabeled examples. Results obtained on very high resolution (VHR) multispectral and hyperspectral images show the relevance of the method in the context of urban image classification. Also, its simplicity and the few parameters involved make the method versatile and workable by unexperienced users.