Mining Land Cover Information Using Multilayer Perceptron and Decision Tree from MODIS Data

Land cover (LC) changes play a major role in global as well as at regional scale patterns of the climate and biogeochemistry of the Earth system. LC information presents critical insights in understanding of Earth surface phenomena, particularly useful when obtained synoptically from remote sensing data. However, for developing countries and those with large geographical extent, regular LC mapping is prohibitive with data from commercial sensors (high cost factor) of limited spatial coverage (low temporal resolution and band swath). In this context, free MODIS data with good spectro-temporal resolution meet the purpose. LC mapping from these data has continuously evolved with advances in classification algorithms. This paper presents a comparative study of two robust data mining techniques, the multilayer perceptron (MLP) and decision tree (DT) on different products of MODIS data corresponding to Kolar district, Karnataka, India. The MODIS classified images when compared at three different spatial scales (at district level, taluk level and pixel level) shows that MLP based classification on minimum noise fraction components on MODIS 36 bands provide the most accurate LC mapping with 86% accuracy, while DT on MODIS 36 bands principal components leads to less accurate classification (69%).

Performance of the generalized delta rule in structural damage detection

The paper examines the suitability of the generalized data rule in training artificial neural networks (ANN) for damage identification in structures. Several multilayer perceptron architectures are investigated for a typical bridge truss structure with simulated damage stares generated randomly. The training samples have been generated in terms of measurable structural parameters (displacements and strains) at suitable selected locations in the structure. Issues related to the performance of the network with reference to hidden layers and hidden. neurons are examined. Some heuristics are proposed for the design of neural networks for damage identification in structures. These are further supported by an investigation conducted on five other bridge truss configurations.

A phase field study of morphological instabilities in multilayer thin films

We studied the microstructural evolution of multiple layers of elastically stiff films embedded in an elastically soft matrix using a phase field model. The coherent and planar film/matrix interfaces are rendered unstable by the elastic stresses due to a lattice parameter mismatch between the film and matrix phases, resulting in the break-up of the films into particles. With an increasing volume fraction of the stiff phase, the elastic interactions between neighbouring layers lead to: (i) interlayer correlations from an early stage; (ii) a longer wavelength for the maximally growing wave; and therefore (iii) a delayed break-LIP. Further, they promote a crossover in the mode of instability from a predominantly anti-symmetric (in phase) one to a symmetric (out of phase) one. We have computed a stability diagram for the most probable mode of break-up in terms of elastic modulus Mismatch and Volume fraction. We rationalize our results in terms of the initial driving force for destabilization, and corroborate our conclusions using simulations in elastically anisotropic systems.

Mean-field results of a lattice-gas model of multilayer adsorption

A lattice-gas model of multilayer adsorption has been solved in the mean-field approximation by a different numerical method. Earlier workers obtained a single solution for all values of temperature and pressure. In the present work, multiple solutions have been obtained in certain regions of temperature and pressure which give rise to bysteresis in the adsorption isotherm. In addition, we have obtained a parameter which behaves like an order parameter for the transition. The potential-energy function shows a double minimum in the region of bysteresis and a single maximum elsewhere.

Ultralow noise field-effect transistor from multilayer graphene

We present low-frequency electrical resistance fluctuations, or noise, in graphene-based field-effect devices with varying number of layers. In single-layer devices, the noise magnitude decreases with increasing carrier density, which behaved oppositely in the devices with two or larger number of layers accompanied by a suppression in noise magnitude by more than two orders in the latter case. This behavior can be explained from the influence of external electric field on graphene band structure, and provides a simple transport-based route to isolate single-layer graphene devices from those with multiple layers. ©2009 American Institute of Physics

LbL Fabricated Poly(Styrene Sulfonate)/TiO2 Multilayer Thin Films for Environmental Applications

Fabrication of multilayer ultrathin composite films composed of nanosized titanium dioxide particles (P25, Degussa) and polyelectrolytes (PELs), such as poly(allyl amine hydrochloride) (PAH) and poly(styrene sulfonate sodium salt) (PSS), on glass substrates using the layer-by-layer (LbL) assembly technique and its potentia application for the photodegradation of rhodamine B under ultraviolet (UV) irradiation has been reported. The polyelectrolytes and TiO2 were deposited on glass substrates at pH 2.5 and the growth of the multilayers was studied using UV/vis speccrophotometer. Thicknes measurements of the films showed a linear increase in film thickness with increase in number of bilayers. The surface microstructure of the thin films was characterized by field emission scanning electron microscope. The ability of the catalysts immobilized by this technique was compared with TiO2 films prepared by drop casting and spin coating methods. Comparison has been made in terms of film stability and photodegradation of rhodamine B. Process variables such as the effect of surface area of the multilayers, umber of bilayers, and initial dye concentration on photodegradation of rhodamine B were studied. Degradation efficiency increased with increase in number of catalysts (total surface area) and bilayers. Kinetics analysis indicated that the photodegradation rates follow first order kinetics. Under maximum loading of TiO2, with five catalyst slides having 20 bilayers of polyelectrolyte/TiO2 on each, 100 mL of 10 mg/L dye solution could be degraded completely in 4 h. The same slides could be reused with the same efficiency for several cycles. This study demonstrates that nanoparticles can be used in wastewater treatment using a simple immobilization technique. This makes the process an attractive option for scale up.

Multilayer single-component thin films and microcapsules via covalent bonded layer-by-layer self-assembly

Fabrication of single-component multilayer thin films still remains a challenging task via the layer-by-layer (LbL) approach. In this communication, we report the self-assembly of single-component multilayer thin films on flat and colloidal substrates through glutaraldehyde mediated covalent bonding.

Random sequential multilayer deposition of different‐sized k mers on a one dimensional infinite substrate

Kinetics of random sequential, irreversible multilayer deposition of macromolecules of two different sizes on a one dimensional infinite lattice is analyzed at the mean field level. A formal solution for the corresponding rate equation is obtained. The Jamming limits and the distribution of gaps of exact sizes are discussed. In the absence of screening, the jamming limits are shown to be the same for all the layers. A detailed analysis for the components differing by one monomer unit is presented. The small and large time behaviors and the dependence of the individual jamming limits of the k mers and (k−1) mers on k and the rate parameters are analyzed.

Multilayer Bi1.5Zn1.0Nb1.5O7/Ba0.6Sr0.4TiO3/Bi1.5Zn1.0Nb1.5O7 thin films for tunable microwave applications

Bi1.5Zn1.0Nb1.5O7/Ba0.6Sr0.4TiO3/Bi1.5Zn1.0Nb1.5O7 tunable multilayer thin film has been fabricated by pulsed laser ablation and characterized. Phase composition and microstructure of multilayer films were characterized by X-ray diffraction, scanning electron microscopy (SEM) and atomic force microscopy (AFM). The film has very smooth surface with RMS roughness of 1.5-2nm and grain size of 100-150 nm. Total film thickness has been measure to be 375 nm. The BZN thin films at 300 K, on Pt(1 1 1)/SiO2/Si substrate showed zero-field dielectric constant of 105 and dielectric loss tangent of 0.002 at frequency of 0.1 MHz. Thin films annealed at 700 degrees C shows the dielectric tunability of 18% with biasing field 500 kV/cm at 0.1 MHz. The multilayer thin film shows nonferroelectric behavior at room temperature. The good physical and electrical properties of multilayer thin films make them promising candidate for tunable microwave device applications. (C) 2010 Elsevier B.V. All rights reserved.

Structural, dielectric and ferroelectric properties of multilayer lithium tantalate thin films prepared by sol-gel technique

Multilayer lithium tantalate thin films were deposited on Pt-Si Si(111)/SiO2/TiO2/Pt(111)]substrates by sol-gel process. The films were annealed at different annealing temperatures (300, 450 and 650 degrees C) for 15 min. The films are polycrystalline at 650 degrees C and at other annealing conditions below 650 degrees C the films are in amorphous state. The films were characterized using X-ray diffraction, atomic force microscopy (AFM) and Raman spectroscopy. The AFM of images show the formation of nanograins of uniform size (50 nm) at 650 degrees C. These polycrystalline films exhibit spontaneous polarization of 1.5 mu C/cm(2) at an application of 100 kV/cm. The dielectric constant of multilayer film is very small (6.4 at 10 kHz) as compared to that of single crystal. (C) 2010 Elsevier B.V. All rights reserved.

Pulsed laser deposited ZnO/ZnO:Li multilayer for blue light emitting diodes

Thin films of ZnO, Li doped ZnO (ZLO) and multilayer of ZnO and ZLO (ZnO/ZLO) were grown on silicon and corning glass substrates by pulsed laser deposition technique. Single phase formation and the crystalline qualities of the films were analyzed by X-ray diffraction and Li composition in the film was investigated to be 15 wt% by X-ray photoelectron spectroscopy. Raman spectrum reveals the hexagonal wurtzite structure of ZnO, ZLO and ZnO/ZLO multilayer and confirms the single phase formation. Films grown on corning glass shows more than 80% transmittance in the visible region and the optical band gaps were calculated to be 3.245, 3.26 and 3.22 eV for ZnO, ZLO and ZnO/ZLO, respectively. An efficient blue emission was observed in all films which were grown on silicon (1 0 0) substrate by photoluminescence (PL). PL measurements at different temperatures reveal that the PL emission intensity of ZnO/ZLO multilayer was weakly dependent on temperature as compared to the single layers of ZnO and ZLO and the wavelength of emission was independent of temperature. Our results indicate that ZnO/ZLO multilayer can be used for the fabrication of blue light emitting diodes. (C) 2011 Elsevier B.V. All rights reserved.

A Multi-layer Perceptron based Non-linear Mixture Model to estimate class abundance from mixed pixels

Sub-pixel classification is essential for the successful description of many land cover (LC) features with spatial resolution less than the size of the image pixels. A commonly used approach for sub-pixel classification is linear mixture models (LMM). Even though, LMM have shown acceptable results, pragmatically, linear mixtures do not exist. A non-linear mixture model, therefore, may better describe the resultant mixture spectra for endmember (pure pixel) distribution. In this paper, we propose a new methodology for inferring LC fractions by a process called automatic linear-nonlinear mixture model (AL-NLMM). AL-NLMM is a three step process where the endmembers are first derived from an automated algorithm. These endmembers are used by the LMM in the second step that provides abundance estimation in a linear fashion. Finally, the abundance values along with the training samples representing the actual proportions are fed to multi-layer perceptron (MLP) architecture as input to train the neurons which further refines the abundance estimates to account for the non-linear nature of the mixing classes of interest. AL-NLMM is validated on computer simulated hyperspectral data of 200 bands. Validation of the output showed overall RMSE of 0.0089±0.0022 with LMM and 0.0030±0.0001 with the MLP based AL-NLMM, when compared to actual class proportions indicating that individual class abundances obtained from AL-NLMM are very close to the real observations.