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.     

Fault diagnosis using partial syndromes: a modified Hopfield neural network approach

This paper presents a modified Hopfield neural network (HNN) for solving the system-level fault diagnosis problem which aims at identifying the set of faulty nodes. This problem has been extensively studied in the last three decades. Nevertheless, identifying the set of all faulty nodes using only partial syndromes, i.e. when some of the testing or comparison outcomes are missing prior to initiating the diagnosis phase, remains an outstanding research issue. The new HNN-based diagnosis algorithm does not require any prior learning or knowledge about the system, nor about any faulty situation, hence providing a better generalisation performance. Results from a thorough simulation study demonstrate the effectiveness of the HNN-based fault diagnosis algorithm in terms of diagnosis correctness, diagnosis latency and diagnosis scalability, for randomly generated diagnosable systems of different sizes and under various fault scenarios. We have also conducted extensive simulations using partial syndromes. Simulations showed that the HNN-based diagnosis performed efficiently, i.e. diagnosis correctness was around 99% when at most half of the test or comparison outcomes are missing, making it a viable alternative to existing diagnosis algorithms.     

Effect of extraction methods on kinetic,chemical composition and antibacterial activities of Tunisian Thymus vulgaris. L. essential oil

The present study aims to compare two innovative extraction techniques: microwave-assisted hydrodistillation (MAHD) and solvent-free microwave extraction (SFME) through traditional extraction techniques: hydrodistillation (HD) and steam distillation (SD) for their efficiency in the extraction of the volatile compounds from Tunisian Thymus vulgaris leaves; the kinetic, yield, composition and antibacterial activities of the essential oil were assessed in vitro. Results show that the essential oils extracted by microwaves were quantitatively (yield) similar to those obtained through the conventional methods, but qualitatively, essential oils analysed by gas chromatography–mass spectrometry (GC–MS) presented 17, 11, 11 and 8 compounds obtained through SFME, MAHD, SD and HD, respectively, mostly consisting of carvacrol (89.24–41.17%), followed by γ-terpinene (11.37–1.37%) and para-cymene (27.95–2.05%). The essential oils were screened for antibacterial activity against 5 microorganisms. All essential oils obtained by studied extraction methods showed the same resistance against Gram (?) and Gram (+) bacteria. The SFME method gave the best results: rapid kinetic of extraction (30 min vs. 35 min for MAHD, 120 min for SD, and 180 min for HD), less energy saving and cleanest process.     

Osmotic Dehydration Kinetics of Pomegranate Seeds Using Date Juice as an Immersion Solution Base

Pomegranate seeds were osmodehydrated using date juice added with sucrose (final °Brix, 55) as immersion solution. The kinetics of osmotic dehydration showed that the most significant changes of mass transfer took place during the first 20 min of the process, regardless of date juice varieties. During this time, seed water loss and solid gain were estimated to be ∼39% and ∼6%, respectively. After 20 min of the process, the percentage of water loss and solid gain varied slightly and ranged on average close to ∼40% and ∼9%, respectively. During osmotic dehydration, there was a leaching of natural solutes from seeds into the solution, which is quantitatively not negligible, and might have an important impact on the sensorial and nutritional value of seeds and date juices. Both scanning electron microscopy and texture (compression) analysis revealed that osmotic dehydration process induced modifications of seed texture and cell structure. Sucrose was found to be the essential element which influences the texture of seed and the viscosity of date juice. Additionally, natural sugar present in date juice permits substituting 35% of the total quantity of sucrose added to the osmotic solution.     

Determination of the mixing performance of sulzer SMV static mixers by laser induced fluorescence

Local concentrations are measured at the outlet of different configurations of Sulzer SMV static mixers placed in a tube. Laser Induced Fluorescence is the non intrusive technique used to measure mean and RMS concentrations. The two fluids mixed are miscible and the flow regime is turbulent. The influence of the number of elements, their position and the ratio between the velocities of the two compounds on the mixing is studied. The results show that these parameters are important for the quality of mixing. The velocity ratio influences the radial mixing, while the mixing length depends on the number and position of the elements.     

A new convolution approach for the time-delay identification of systems with arbitrary entries

In this paper, an online delay identification approach of continuous-time linear systems with unstructured entries is achieved via a new algebraic technique. The arbitrary input and output trajectories are chosen with close and abundant crossing zero. Initial conditions and static disturbances are taken into account in the design of the identification approach. The proposed method is based on a distributional algebraic technique and a convolution approach. A proposed theorem is hence enounced to identify a single time-delay of such systems. The effectiveness of the proposed approach is demonstrated by an illustrative example. The obtained results show the high performances of the proposed time-delay identification approach in severe operation conditions of the considered system.     

Utilisation of attenuated total reflectance MIR and front-face fluorescence spectroscopies for the identification of Saint-Nectaire cheeses varying by manufacturing conditions

Twelve samples of cheeses (three types of Saint-Nectaire PDO cheeses and Savaron cheeses) differing by manufacturing and ripening conditions, from 12 different producers, were characterised by attenuated total reflectance mid-infrared (MIR) and front-face fluorescence spectroscopies, dynamic testing rheology and physico-chemical analysis. Fluorescence spectra (tryptophan residues, vitamin A and riboflavin) and MIR (3,000–2,800 (fat region), 1,700–1,500 (protein region) and 1,500–900 cm⁻¹ (fingerprint region)) spectra were recorded on cheese samples. The potential of the data tables was investigated for discriminating the four different groups of cheeses. The results of factorial discriminant analysis (FDA) performed on the fluorescence and mid-infrared spectra showed a good discrimination of the four cheese groups. Considering cross-validation results, the best classifications (100%) were achieved from mid-infrared and fluorescence spectra, while only 91.7 and 72.2% of correct classification were obtained by applying FDA to rheology and physico-chemical data, respectively. Spectroscopic techniques could provide useful fingerprints and allow the identification of investigated cheeses according to manufacturing conditions. Simple and rapid spectroscopic methods offer a promising approach to the authentication of cheeses.     

Prediction of the rheology parameters of ripened semi-hard cheeses using fluorescence spectra in the UV and visible ranges recorded at a young stage

Storage modulus (G′), loss modulus (G″), strain, tan (δ) and complex viscosity (η*) of 20 semi-hard cheeses were measured by dynamic oscillatory analysis after 2, 30 and 60 days of ripening. On the same cheeses and at the same ages, tryptophan and riboflavin fluorescence spectra were recorded. The aim was to predict the rheology parameters of ripened cheeses from spectra recorded on these cheeses at a young stage. Using partial least square, tryptophan fluorescence spectra recorded at 20 °C on 2-days-old cheeses predicted G′, G″, strain, tan (δ) and η* measured at 80 °C on the 60-days-old cheeses with correlation coefficients (R) of 0.98, 0.97, 0.98, 0.98 and 0.97, respectively. Riboflavin fluorescence spectra gave slightly lower correlation coefficients of 0.88, 0.88, 0.92, 0.87 and 0.88, respectively. Dependent only on visible light, the riboflavin fluorescence spectra potentially provide viable and economic prediction of the rheology of ripe cheese.     

An evolutionary algorithm for abductive reasoning

Abductive reasoning (or abduction) is the process of inferring hypotheses from observed data using a certain ‘knowledge’ encoded in the form of inference rules (or causal relations). Many important kinds of intellectual tasks, including medical diagnosis, fault diagnosis, scientific discovery, legal reasoning, and natural language understanding have been characterised as abduction. Unfortunately, abduction is 𝒩𝒫-hard. Genetic algorithms and biologically motivated computational paradigms inspired by the natural evolution turned out to be efficient in solving many hard problems while other existing approaches failed to solve in general. In this article, we present a genetic algorithm called HAKIM, for solving abduction problems. We encode an explanation in a chromosome-like structure, where every gene models a possible single hypothesis. Thereafter, we develop a fitness function that characterises the overall ‘quality’ of a chromosome representing an explanation; and then use standard genetic operators to compute a set of hypotheses that best explains the observed data. Simulation results on large-scale medical problems reveal the good performance of our model HAKIM.     

Prediction of sensory attributes of European Emmental cheese using near-infrared spectroscopy: A feasibility study

The present work evaluated the ability of near infrared (NIR) spectroscopy in predicting some sensory attributes of 20 Emmental cheeses originating from different European regions. For the purpose of this study four appearance and texture attributes, namely, adhesivity, friability, elasticity and firmness and six olfacto-gustatory attributes namely, aroma intensity, odour intensity, bitterness, saltiness, acidity and sweetness were selected by the sensory panel. Calibration models between sensory properties and NIR spectra were developed using partial least squares (PLS) regression. The squared correlation coefficients (R²) were greater than 0.5 for adhesivity, elasticity, firmness, aroma, bitterness, saltiness, acidity and sweetness. In addition, a good correlation between sensory attributes and NIR spectra was found using canonical correlation analysis (CCA). Therefore, this work demonstrates the feasibility of NIR to predict some sensory attributes since a relatively high correlation between sensory data and NIR spectra was found. However, further research with a large data bases will be needed in order to validate the method.