This paper presents two semisupervised one-class support vector machine (OC-SVM) classifiers for remote sensing applications. In one-class image classification, one tries to detect pixels belonging ...to one of the classes in the image and reject the others. When few labeled pixels of only one class are available, obtaining a reliable classifier is a difficult task. In the particular case of SVM-based classifiers, this task is even harder because the free parameters of the model need to be finely adjusted, but no clear criterion can be adopted. In order to improve the OC-SVM classifier accuracy and alleviate the problem of free-parameter selection, the information provided by unlabeled samples present in the scene can be used. In this paper, we present two state-of-the-art algorithms for semisupervised one-class classification for remote sensing classification problems. The first proposed algorithm is based on modifying the OC-SVM kernel by modeling the data marginal distribution with the graph Laplacian built with both labeled and unlabeled samples. The second one is based on a simple modification of the standard SVM cost function which penalizes more the errors made when classifying samples of the target class. The good performance of the proposed methods is illustrated in four challenging remote sensing image classification scenarios where the goal is to detect one of the classes present on the scene. In particular, we present results for multisource urban monitoring, hyperspectral crop detection, multispectral cloud screening, and change-detection problems. Experimental results show the suitability of the proposed techniques, particularly in cases with few or poorly representative labeled samples.
ESA's upcoming satellites Sentinel-2 (S2) and Sentinel-3 (S3) aim to ensure continuity for Landsat 5/7, SPOT-5, SPOT-Vegetation and Envisat MERIS observations by providing superspectral images of ...high spatial and temporal resolution. S2 and S3 will deliver near real-time operational products with a high accuracy for land monitoring. This unprecedented data availability leads to an urgent need for developing robust and accurate retrieval methods. Machine learning regression algorithms may be powerful candidates for the estimation of biophysical parameters from satellite reflectance measurements because of their ability to perform adaptive, nonlinear data fitting.
By using data from the ESA-led field campaign SPARC (Barrax, Spain) we have compared the utility of four state-of-the-art machine learning regression algorithms and four different S2 and S3 band settings to assess three important biophysical parameters: leaf chlorophyll content (Chl), leaf area index (LAI) and fractional vegetation cover (FVC). The tested Sentinel configurations were: S2-10m (4 bands), S2-20m (8 bands), S2-60m (10 bands) and S3-300m (19 bands), and the tested methods were: neural networks (NN), support vector regression (SVR), kernel ridge regression (KRR), and Gaussian processes regression (GPR).
GPR outperformed the other retrieval methods for the majority of tested configurations and was the only method that reached the 10% precision required by end users in the estimation of Chl. Also, although validated with an RMSE accuracy around 20%, GPR yielded optimal LAI and FVC estimates at highest S2 spatial resolution of 10m with only four bands. In addition to high accuracy values, GPR also provided confidence intervals of the estimates and insight in relevant bands, which are key advantages over the other methods. Given all this, GPR proved to be a fast and accurate nonlinear retrieval algorithm that can be potentially implemented for operational monitoring applications.
► Sentinel-2 and -3 aim to ensure continuity of operational land monitoring products. ► Machine learning algorithms are promising for biophysical parameters retrieval. ► Gaussian Processes regression (GPR) is evaluated as best performing algorithm. ► GPR reached accuracies superior to the by GMES requested threshold of 10%. ► Along with estimates, GPR provides insight in relevant bands and confidence values.
This paper presents a semi-supervised graph-based method for the classification of hyperspectral images. The method is designed to handle the special characteristics of hyperspectral images, namely, ...high-input dimension of pixels, low number of labeled samples, and spatial variability of the spectral signature. To alleviate these problems, the method incorporates three ingredients, respectively. First, being a kernel-based method, it combats the curse of dimensionality efficiently. Second, following a semi-supervised approach, it exploits the wealth of unlabeled samples in the image, and naturally gives relative importance to the labeled ones through a graph-based methodology. Finally, it incorporates contextual information through a full family of composite kernels. Noting that the graph method relies on inverting a huge kernel matrix formed by both labeled and unlabeled samples, we originally introduce the Nystro umlm method in the formulation to speed up the classification process. The presented semi-supervised-graph-based method is compared to state-of-the-art support vector machines in the classification of hyperspectral data. The proposed method produces better classification maps, which capture the intrinsic structure collectively revealed by labeled and unlabeled points. Good and stable accuracy is produced in ill-posed classification problems (high dimensional spaces and low number of labeled samples). In addition, the introduction of the composite-kernel framework drastically improves results, and the new fast formulation ranks almost linearly in the computational cost, rather than cubic as in the original method, thus allowing the use of this method in remote-sensing applications.
Photosynthesis is the process by which plants harvest sunlight to produce sugars from carbon dioxide and water. It is the primary source of energy for all life on Earth; hence it is important to ...understand how this process responds to climate change and human impact. However, model-based estimates of gross primary production (GPP, output from photosynthesis) are highly uncertain, in particular over heavily managed agricultural areas. Recent advances in spectroscopy enable the space-based monitoring of sun-induced chlorophyll fluorescence (SIF) from terrestrial plants. Here we demonstrate that spaceborne SIF retrievals provide a direct measure of the GPP of cropland and grassland ecosystems. Such a strong link with crop photosynthesis is not evident for traditional remotely sensed vegetation indices, nor for more complex carbon cycle models. We use SIF observations to provide a global perspective on agricultural productivity. Our SIF-based crop GPP estimates are 50-75% higher than results from state-of-the-art carbon cycle models over, for example, the US Corn Belt and the Indo-Gangetic Plain, implying that current models severely underestimate the role of management. Our results indicate that SIF data can help us improve our global models for more accurate projections of agricultural productivity and climate impact on crop yields. Extension of our approach to other ecosystems, along with increased observational capabilities for SIF in the near future, holds the prospect of reducing uncertainties in the modeling of the current and future carbon cycle.
Biophysical parameters such as leaf chlorophyll content (LCC) and leaf area index (LAI) are standard vegetation products that can be retrieved from Earth observation imagery. This paper introduces a ...new machine learning regression algorithms (MLRAs) toolbox into the scientific Automated Radiative Transfer Models Operator (ARTMO) software package. ARTMO facilitates retrieval of biophysical parameters from remote observations in a MATLAB graphical user interface (GUI) environment. The MLRA toolbox enables analyzing the predictive power of various MLRAs in a semiautomatic and systematic manner, and applying a selected MLRA to multispectral or hyperspectral imagery for mapping applications. It contains both linear and nonlinear state-of-the-art regression algorithms, in particular linear feature extraction via principal component regression (PCR), partial least squares regression (PLSR), decision trees (DTs), neural networks (NNs), kernel ridge regression (KRR), and Gaussian processes regression (GPR). The performance of multiple implemented regression strategies has been evaluated against the SPARC dataset (Barrax, Spain) and simulated Sentinel-2 (8 bands), CHRIS (62 bands) and HyMap (125 bands) observations. In general, nonlinear regression algorithms (NN, KRR, and GPR) outperformed linear techniques (PCR and PLSR) in terms of accuracy, bias, and robustness. Most robust results along gradients of training/validation partitioning and noise variance were obtained by KRR while GPR delivered most accurate estimations. We applied a GPR model to a hyperspectral HyMap flightline to map LCC and LAI. We exploited the associated uncertainty intervals to gain insight in the per-pixel performance of the model.
A framework for semisupervised remote sensing image classification based on neural networks is presented. The methodology consists of adding a flexible embedding regularizer to the loss function used ...for training neural networks. Training is done using stochastic gradient descent with additional balancing constraints to avoid falling into local minima. The method constitutes a generalization of both supervised and unsupervised methods and can handle millions of unlabeled samples. Therefore, the proposed approach gives rise to an operational classifier, as opposed to previously presented transductive or Laplacian support vector machines (TSVM or LapSVM, respectively). The proposed methodology constitutes a general framework for building computationally efficient semisupervised methods. The method is compared with LapSVM and TSVM in semisupervised scenarios, to SVM in supervised settings, and to online and batch k -means for unsupervised learning. Results demonstrate the improved classification accuracy and scalability of this approach on several hyperspectral image classification problems.
This paper presents the framework of kernel-based methods in the context of hyperspectral image classification, illustrating from a general viewpoint the main characteristics of different ...kernel-based approaches and analyzing their properties in the hyperspectral domain. In particular, we assess performance of regularized radial basis function neural networks (Reg-RBFNN), standard support vector machines (SVMs), kernel Fisher discriminant (KFD) analysis, and regularized AdaBoost (Reg-AB). The novelty of this work consists in: 1) introducing Reg-RBFNN and Reg-AB for hyperspectral image classification; 2) comparing kernel-based methods by taking into account the peculiarities of hyperspectral images; and 3) clarifying their theoretical relationships. To these purposes, we focus on the accuracy of methods when working in noisy environments, high input dimension, and limited training sets. In addition, some other important issues are discussed, such as the sparsity of the solutions, the computational burden, and the capability of the methods to provide outputs that can be directly interpreted as probabilities.
The increase in spatial and spectral resolution of the satellite sensors, along with the shortening of the time-revisiting periods, has provided high-quality data for remote sensing image ...classification. However, the high-dimensional feature space induced by using many heterogeneous information sources precludes the use of simple classifiers: thus, a proper feature selection is required for discarding irrelevant features and adapting the model to the specific problem. This paper proposes to classify the images and simultaneously to learn the relevant features in such high-dimensional scenarios. The proposed method is based on the automatic optimization of a linear combination of kernels dedicated to different meaningful sets of features. Such sets can be groups of bands, contextual or textural features, or bands acquired by different sensors. The combination of kernels is optimized through gradient descent on the support vector machine objective function. Even though the combination is linear, the ranked relevance takes into account the intrinsic nonlinearity of the data through kernels. Since a naive selection of the free parameters of the multiple-kernel method is computationally demanding, we propose an efficient model selection procedure based on the kernel alignment. The result is a weight (learned from the data) for each kernel where both relevant and meaningless image features automatically emerge after training the model. Experiments carried out in multi- and hyperspectral, contextual, and multisource remote sensing data classification confirm the capability of the method in ranking the relevant features and show the computational efficience of the proposed strategy.
This paper analyzes the classification of hyperspectral remote sensing images with linear discriminant analysis (LDA) in the presence of a small ratio between the number of training samples and the ...number of spectral features. In these particular ill-posed problems, a reliable LDA requires one to introduce regularization for problem solving. Nonetheless, in such a challenging scenario, the resulting regularized LDA (RLDA) is highly sensitive to the tuning of the regularization parameter. In this context, we introduce in the remote sensing community an efficient version of the RLDA recently presented by Ye to cope with critical ill-posed problems. In addition, several LDA-based classifiers (i.e., penalized LDA, orthogonal LDA, and uncorrelated LDA) are compared theoretically and experimentally with the standard LDA and the RLDA. Method differences are highlighted through toy examples and are exhaustively tested on several ill-posed problems related to the classification of hyperspectral remote sensing images. Experimental results confirm the effectiveness of the presented RLDA technique and point out the main properties of other analyzed LDA techniques in critical ill-posed hyperspectral image classification problems.
PDF
