Proc. SPIE. 7285, International Conference on Earth Observation Data Processing and Analysis (ICEODPA)
KEYWORDS: Data modeling, Geographic information systems, Artificial neural networks, Nonlinear dynamics, Neural networks, Space operations, Autoregressive models, Performance modeling, Time series analysis, Neurons
Spatio-Temporal Autoregressive Integrated Moving Average (STAIRMA) model family is a very useful tool in
modeling space-time series data. It assumes that space-time series data is correlated linearly in space and time. However,
in reality most space-time series contains nonlinear space-time autocorrelation structure, which can't be modeled by
STARIMA. Artificial neural networks (ANN) have shown great flexibility in modeling and forecasting nonlinear
dynamic process. In the paper, we developed an architecture approach to model space-time series data using artificial
neural network (ANN). The model is tested with forest fire prediction in Canada. The experimental result demonstrates
that STANN achieves much better prediction accuracy than STARIMA model.
ZnO:Al is a kind of N type semiconductor material with low resistance and high transmittance in the visible region.
Using ZnO mixed with Al<sub>2</sub>O<sub>3</sub> (2 wt %) as target, ZAO thin films were deposited on glass substrate by RF magnetron
sputtering. Orthogonal experiments were used to analyze the effects of main factors (oxygen flux, argon pressure,
substrate temperature, RF power) on the properties (transmittance, resistance) of the film. The results showed that the
optimal parameters in the room temperature are: the partial pressure of argon without oxygen is 0.1 Pa, RF power is
400w. After vacuum annealing at 220°C, the deposited film exhibits visible transmittance of above 82% and minimum
sheet resistance in 3.36 × 10<sup>-3</sup> Ω • cm.
During optical coatings monitoring, the test glass edgy-effect often makes the monitoring curve falseness and thin films
thickness control inaccurate. The thin films edge-effect makes the films thickness different from the test glass centre to
the edge, and the edge-effect gets distinctly with coating layers increasing. NBBF (narrow band pass filter) is fabricated,
and its monitoring curve and spectrum curve are analyzed. The results show that the edge-effect comes from material
deposition angle when test glass does not rotate, temperature and electric field different distributing on the test glass
surface. Several methods are used to minish the test glass edge-effect, such as, rotating the test glass to reduce the films
thickness difference caused by material deposition angle, using quartz and other glass alike material as a link between
test glass and the fixture to lessen temperature and electric field different distributing, making beam size small monitor
the test glass centre field, where can be considered having no thickness different. The above methods make the thickness
symmetrical over the test glass, and then the experiment monitoring curve is close to the theory curve. The results are
important for the thin films automatic monitoring, especially for NBBF coatings.
By introducing structure perturbation coefficients with different levels, relevant effects on PBG and
density of mode (DOM) in 1D photonic crystal had been detailedly discussed under specifying numbers of bilayer
and contrast of refractive index in this paper. Numerical simulation indicates that there are remarkable influences
on optical properties of 1D photonic crystal resulted by structure perturbation, especially for PBG location and
DOM at band edge. Generally, both PBG shifting and shrinking occurs due to disruptive periodicity. In specific
case, especially along with perturbation coefficient increasing, it is found that PBG is extended obviously and the
higher DOM at the band edge also can be obtained. According to this, some conclusions had been drawn which are
significant to developing omnidirectional reflectors, band edge lasers and other devices based 1D photonic crystals.