Moussa Sofiane Karoui

Blind spatial unmixing of multispectral images: New methods combining sparse component analysis, clustering and non-negativity constraints

Remote sensing has become an unavoidable tool for better managing our environment, generally by realizing maps of land cover using classification techniques. Traditional classification techniques assign only one class (e.g., water, soil, grass) to each pixel of remote sensing images. However, the area covered by one pixel contains more than one surface component and results in the mixture of these surface components. In such situations, classical classification is not acceptable for many major applications, such as environmental monitoring, agriculture, mineral exploration and mining, etc. Most methods proposed for treating this problem have been developed for hyperspectral images. On the contrary, there are very few automatic techniques suited to multispectral images. In this paper, we propose new unsupervised spatial methods (called 2D-Corr-NLS and 2D-Corr-NMF) in order to unmix each pixel of a multispectral image for better recognizing the surface components constituting the observed scene. These methods are related to the blind source separation (BSS) problem, and are based on sparse component analysis (SCA), clustering and non-negativity constraints. Our approach consists in first identifying the mixing matrix involved in this BSS problem, by using the first stage of a spatial correlation-based SCA method with very limited source sparsity constraints, combined with clustering. Non-negative least squares (NLS) or non-negative matrix factorization (NMF) methods are then used to extract spatial sources. An important advantage of our proposed methods is their applicability to the possibly globally underdetermined, but locally (over)determined BSS model in multispectral remote sensing images. Experiments based on realistic synthetic mixtures and real multispectral images collected by the Landsat ETM+ and the Formosat-2 sensors are performed to evaluate the performance of the proposed approach. We also show that our methods significantly outperform the sequential maximum angle convex cone (SMACC) method.

Improvement of remote sensing multispectral image classification by using Independent Component Analysis

This paper deals with the application of Independent Component Analysis (ICA) as a solution to Blind Source Separation (BSS), in order to pre-process remote sensing multispectral images before we classify them. We analyze the structure of the considered data, and especially show that each recorded image corresponding to a spectral band may be seen as an observation consisting of a mixture (linear combination) of source images. The latter images correspond to the abundances of the pure elements (endmembers) in the pixels. Using BSS methods, one can hope to reduce the mixing effect in these observations, which then allows better recognition of the classes constituting the observed scene. Based on this approach, we create new images (i.e. at least partly separated images) by using ICA, starting from HRV SPOT images. These images are then used as inputs of a supervised classifier integrating textural information. The separated image classification results show a clear improvement compared to classification of initial images. This show the contribution of ICA as an attractive pre-processing for classification of multispectral remote sensing imagery.

Joint nonnegative matrix factorization for hyperspectral and multispectral remote sensing data fusion

This paper presents a new fusion approach producing unobservable fused remote sensing data with high spatial and spectral resolutions. This approach, related to linear spectral unmixing (LSU) techniques, introduces joint nonnegative matrix factorization (JNMF) for combining observable low spatial resolution hyperspectral and high spatial resolution multispectral data. JNMF is applied to synthetic but realistic data generated from real airborne hyperspectral data. Spectral and spatial qualities of fused data are evaluated by frequently used criteria. Experimental results show the low computational cost of the proposed approach, and the good spectral and spatial fidelities of the fused data. Our method also outperforms the recently proposed coupled nonnegative matrix factorization (CNMF) method.

A new spatial sparsity-based method for extracting endmember spectra from hyperspectral data with some pure pixels

Remote sensing hyperspectral sensors typically collect data in contiguous narrow bands (up to several hundred bands) in the electromagnetic spectrum. In hyperspectral imagery, pixels are often linear mixtures of pure materials (endmembers) contained in the observed scene. In this paper, we propose a new unsupervised spatial method (called 2D-VM) for endmember spectra extraction from data to be collected by future higher spatial resolution hyperspectral sensors, which will allow the existence of some pure pixels. This method is related to the Blind Mixture Identification (BMI) problem, and is based on Sparse Component Analysis (SCA). It extracts the endmember spectra by using a spatial variance-based SCA method, which detects a few pure-pixel zones. Experiments based on synthetic but realistic data are performed to compare the performance of the proposed approach and of methods from the literature. We show that our approach outperforms all other methods.

Hypersharpening by Joint-Criterion Nonnegative Matrix Factorization

Hypersharpening aims at combining an observable low-spatial resolution hyperspectral image with a high-spatial resolution remote sensing image, in particular a multispectral one, to generate an unobservable image with the high spectral resolution of the former and the high spatial resolution of the latter. In this paper, two such new fusion methods are proposed. These methods, related to linear spectral unmixing techniques, and based on nonnegative matrix factorization (NMF), optimize a new joint criterion and extend the recently proposed joint NMF (JNMF) method. The first approach, called gradient-based joint-criterion NMF (Grd-JCNMF), is a gradient-based method. The second one, called multiplicative JCNMF (Mult-JCNMF), uses new designed multiplicative update rules. These two JCNMF approaches are applied to synthetic and semireal data, and their effectiveness, in spatial and spectral domains, is evaluated with commonly used performance criteria. Experimental results show that the proposed JCNMF methods yield sharpened hyperspectral data with good spectral and spatial fidelities. The obtained results are compared with the performance of two NMF-based methods and one approach based on a sparse representation. These results show that the proposed methods significantly outperform the well-known coupled NMF sharpening method for most performance figures. Also, the proposed Mult-JCNMF method provides the results that are similar to those obtained by JNMF, with a lower computational cost. Compared with the tested sparse-representation-based approach, the proposed methods give better results. Moreover, the proposed Grd-JCNMF method considerably surpasses all other tested methods.

Pansharpening multispectral remote sensing data by multiplicative joint nonnegative matrix factorization

ABSTRACT Pansharpening aims at combining observable panchromatic and multispectral images to generate an unobservable image with the high spatial resolution of the former and the spectral diversity of the latter. In this paper a new fusion method is proposed. This method, related to linear spectral unmixing (LSU) techniques and based on non-negative matrix factorization (NMF), optimizes, by new iterative–multiplicative update rules, a joint criterion that exploits a spatial degradation model between the two images. The proposed Multiplicative Joint Non-negative Matrix Factorization (MJNMF) approach is applied to synthetic and real data, and its effectiveness in spatial and spectral domains is evaluated with commonly used performance criteria. Experimental results show that the proposed method yields good spectral and spatial fidelities of the pansharpened data. Also, it outperforms those tested from the literature.

Hyperspectral data multi-sharpening based on linear-quadratic nonnegative matrix factorization

In this paper, we propose a new multi-sharpening approach for improving the spatial resolution of hyperspectral data. This approach, based on the linear-quadratic spectral unmixing concept, uses a linear-quadratic nonnegative matrix factorization multiplicative algorithm. Our method first consists in unmixing the low spatial resolution hyperspectral data and high spatial resolution multispectral data. The obtained high resolution spectral and spatial parts of information are then recombined, according to the linear-quadratic mixing model, in order to obtain unobservable multi-sharpened high spatial resolution hyperspectral data. Experiments, based on realistic synthetic and real data, are carried out to evaluate the performance of the proposed approach and of linear nonnegative matrix factorization-based approaches from the literature. We show that our proposed approach significantly outperforms the used literature methods.

Hyperspectral image unmixing by non-negative matrix factorization initialized with modified independent component analysis

In this paper, we propose an unsupervised unmixing approach for hyperspectral images, consisting of a modified version of ICA, followed by NMF. In the ideal case of a hyperspectral image combining (C-1) statistically independent source images, and a Cth image which is dependent on them due to the sum-to-one constraint, our modified ICA first estimates these (C-1) sources and associated mixing coefficients, and then derives the remaining source and coefficients, while also removing the BSS scale indeterminacy. In real conditions, the above (C-1) sources may be somewhat dependent. Our modified ICA method then only yields approximate data. These are then used as the initial values of an NMF method, which refines them. Our tests show that this joint modifICA-NMF approach significantly outperforms the considered classical methods.

Classification of Quickbird imagery over urban area using convolutional neural network

During the past decades significant efforts have been made in developing various methods for Very high spatial resolution (VHSR) remotely sensed image classification; most of them are based on handcrafted learning-based features. Recently deep learning-based techniques have demonstrated excellent performance in remote sensing applications. In this paper we address the problem of urban imagery classification by developing a convolutional neural network (CNN) approach, which are the most popular deep learning approach for image classification. We design a custom CNN that operates on local patches in order to produce pixel-level classification map. The performance of the proposed model is validated on an exhaustive experimental comparison on a set of 20 QuickBird pansharpened multi-spectral images in urban zones. The obtained results outperform those obtained by different classification approaches on the same dataset.

Modified nonnegative matrix factorization for endmember spectra extraction from highly mixed hyperspectral images combined with multispectral data

In this paper, a new approach is proposed for linear endmember spectra extraction from a highly mixed hyperspectral image combined with high spatial resolution multispectral data containing pure pixels. This new approach, which is applied to unmix the considered hyperspectral image, is based on a modified version of nonnegative matrix factorization (NMF) coupled with nonnegative least squares (NLS). The multispectral data are used to initialize the hyperspectral NMF algorithm and to constrain it during matrix updates. Experiments based on synthetic and real data are performed to evaluate the performance of the proposed approach and to compare it with five methods from the literature only applied to the hyperspectral data. The obtained performance shows the superiority of the proposed approach as compared with all other methods. Also, the impact, on the proposed method, of spectral variability between hyperspectral and multispectral data is evaluated, and the obtained results show the robustness of the proposed method to this variability.