Laurence Hubert-Moy

Exploring subtle land use and land cover changes: a framework for future landscape studies

Land cover and land use changes can have a wide variety of ecological effects, including significant impacts on soils and water quality. In rural areas, even subtle changes in farming practices can affect landscape features and functions, and consequently the environment. Fine-scale analyses have to be performed to better understand the land cover change processes. At the same time, models of land cover change have to be developed in order to anticipate where changes are more likely to occur next. Such predictive information is essential to propose and implement sustainable and efficient environmental policies. Future landscape studies can provide a framework to forecast how land use and land cover changes is likely to react differently to subtle changes. This paper proposes a four step framework to forecast landscape futures at fine scales by coupling scenarios and landscape modelling approaches. This methodology has been tested on two contrasting agricultural landscapes located in the United States and France, to identify possible landscape changes based on forecasting and backcasting agriculture intensification scenarios. Both examples demonstrate that relatively subtle land cover and land use changes can have a large impact on future landscapes. Results highlight how such subtle changes have to be considered in term of quantity, location, and frequency of land use and land cover to appropriately assess environmental impacts on water pollution (France) and soil erosion (US). The results highlight opportunities for improvements in landscape modelling.

A Comparison of Parametric Classification Procedures of Remotely Sensed Data Applied on Different Landscape Units

This paper presents an evaluation of several parametric classification algorithms to assess their accuracy on various landscapes. Traditionally the maximum likelihood classifier is used to obtain thematic maps in land use. In this work different classification algorithms including contextual classifiers, one of them being original, are applied and compared on sites belonging to landscape units ranging from tiny fields surrounded by hedges to larger and more open fields. Confusion matrices and result analysis are presented at two observation scales: at the catchment area level and at the landscape unit level. We show how the choice of a classification technique can significantly influence the results of crop inventories and how the accuracy of classification algorithms vary according to the landscape units of the studied area. From these results a strategy can be developed for a better choice of classification algorithms regarding the considered landscape structure.

Independent component analysis as a tool for the dimensionality reduction and the representation of hyperspectral images

Independent component analysis (ICA) is a multivariate data analysis process largely studied these last years in the signal processing community for blind source separation. This paper proposes to show the interest of ICA as a tool for unsupervised analysis of hyperspectral images. The commonly used principal component analysis (PCA) is the mean square optimal projection for gaussian data leading to uncorrelated components by using second order statistics. ICA rather uses higher order statistics and leads to independent components, a stronger statistical assumption revealing interesting features in the usually non gaussian hyperspectral data sets.

Land cover change prediction with a new theory of plausible and paradoxical reasoning

The spatial prediction of land cover ot the Jeld scale in winter appears useful for the issue of bare soils reduction in agricultural intensive regions. High variability of the factors that motivate the land cover changes between each winter involves integration of uncertainty in the modelling process. Fusion process with Dempster-Shafer Theoiy (DST) presents some limits in generating errors in decision making when the degree of conflict, between the sources of evidence that suppor? land cover hypotheses, becomes important. This paper focuses on the application of Dezeri-Smarandache Theory (DSmT) method to the fusion of multiple land-use attributes for land cover prediction purpose. Results are discussed and compared with prediction levels achieved with DST. Through this

Classification of hyperspectral images with nonlinear filtering and support vector machines

rst application of the Dezert- Smarandache Theory, we show an example of this new approach abiliry to solve some of practical problems where the Dempster-Shafer Theory usuolly foils.

Nonlinear filtering of hyperspectral images with anisotropic diffusion

Support vector machines, recently introduced in hyperspectral imagery, are applied to classify land cover on images from the airborne CASI sensor with a small training set. A smoothing preprocessing step is achieved, based on a vectorial extension of the anisotropic diffusion nonlinear filtering process. It allows the separability of the classes to be increased as well as homogeneous areas to be smoothed. It comes to take into consideration the spatial context before the classification, leading to improve the classification rate and to produce noiselessly classification maps with support vector machines.

Combined Use of Multi-Temporal Optical and Radar Satellite Images for Grassland Monitoring

A vectorial extension of the scalar anisotropic diffusion nonlinear filtering process applied on hyperspectral images is presented. In a first step, data are projected in a transformed space with a Maximum Noise Fraction transform, allowing the new components to be sorted in order of signal to noise ratio. The filtering is adapted to the signal to noise ratio of each component and a spectral dissimilarity vectorial measure is used in the filtering process. The inverse transform allows the filtered data to be reprojected in the original space. This process is useful for denoising hyperspectral images and for reducing spatial and spectral variability in each class of interest, leading to increase the performance of further segmentation or classification algorithms.

Spectral unmixing of hyperspectral images with the independent component analysis and wavelet packets

The aim of this study was to assess the ability of optical images, SAR (Synthetic Aperture Radar) images and the combination of both types of data to discriminate between grasslands and crops in agricultural areas where cloud cover is very high most of the time, which restricts the use of visible and near-infrared satellite data. We compared the performances of variables extracted from four optical and five SAR satellite images with high/very high spatial resolutions acquired during the growing season. A vegetation index, namely the NDVI (Normalized Difference Vegetation Index), and two biophysical variables, the LAI (Leaf Area Index) and the fCOVER (fraction of Vegetation Cover) were computed using optical time series and polarization (HH, VV, HV, VH). The polarization ratio and polarimetric decomposition (Freeman–Durden and Cloude–Pottier) were calculated using SAR time series. Then, variables derived from optical, SAR and both types of remotely-sensed data were successively classified using the Support Vector Machine (SVM) technique. The results show that the classification accuracy of SAR variables is higher than those using optical data (0.98 compared to 0.81). They also highlight that the combination of optical and SAR time series data is of prime interest to discriminate grasslands from crops, allowing an improved classification accuracy.

Performance of change detection using remotely sensed data and evidential fusion : comparison of three cases of application

The study addresses the problem of spectral unmixing hyperspectral images, technique allowing the spectra and abundance of each pure material present in each pixel of a scene to be extracted. We first remark that the linear model commonly used in spectral unmixing is exactly the same as the model used in the independant component analysis (ICA), a blind source separation technique studied in the signal processing community; ICA allows each source to be extracted from the observation of some linear combinations-of these ones, based on the assumption of their statistical independence. We show the interest of analyzing the spectra issued from a wavelet packets transformation in order to deal with the assumption of independence, which is usually not verified for natural spectra. A pyramidal algorithm is implemented, allowing the problem of the great number of observations to be addressed.

Evaluation of SPOT imagery for the estimation of grassland biomass

The detection of changes affecting continental surfaces has important applications in hydrological, meteorological and climatic modelling. Using remote sensing data, numerous change indices have already been proposed. Previous work showed the interest of combining several of these to improve change detection performance, using the Dempster–Shafer evidence theory framework. This study analyses the performance of different change indices and their combination in different cases of application: forest logging either in pine forest or in mixed forest, and winter vegetation cover of fields in intensive farming areas, in comparison to the forest fire case presented in previous work. The interest of indices derived from Information Theory, some of which are original, is shown.