- 最后登录
- 2017-9-18
- 注册时间
- 2011-1-12
- 阅读权限
- 90
- 积分
- 12276
- 纳金币
- 5568
- 精华
- 0
|
1 Introduction
This paper proposes a new modeling method for visual simulation
of aurora.The proposed method generates an aurora animation
based on the simulation of the electromagnetic field and emission
principles. The proposed method has two major advantages (1) it
can mimic the phenomena including splitting and rotation, which
are observed in actual aurora, and (2) the color of aurora images
generated by the proposed method closely matches the actual aurora.
The proposed method can create realistic aurora animations.
Baranoski et al. have proposed a method for simulating and rendering
aurora based on the average approximation of the actuallyobserved
data [G. V. G. Baranoski 2003].This method was, however,
limited to static images, and there were some limitations
in the color of the aurora.Baranoski et al. later proposed
another method for modeling aurora-dynamics simulation
[G. V. G. Baranoski 2005].This method can reproduce a rotation
of aurora on thin sheets by simulating forces of electrically
charged particles.This method, however, could not reproduce
a splitting.Ishikawa proposed a method for predicting the region
over the earth where aurora is visible based on the observed
data [Tomokazu Ishikawa 2011].Although this method was effective
for predicting global movement of aurora, it could not generate
a detailed image of aurora.
2 Modeling The Aurora
The proposed method reproduces the aurora movement by simulating
the motions of charged particles.The rest of the paper denotes
y as altitude, and xz plane corresponds to the earth surface.The
proposed method defines the aurora curve, which is a linear combination
of B`ezier and sine curves defined on the xz plane.A width
is also defined at every point on the aurora curve.The aurora curve
with width information is given as input to the proposed method
and is used for calculating the initial positions of the charged particles.
The initial particles are randomly placed above the domain
defined by the aurora curve and the width, and y is 500km which
is the maximum altitude that aurora appears. The proposed method
creates a rectangular grid that encloses all the charged particles on
e-mail:kojima.takafumi0@gmail.com
ye-mail:s2rita@nifty.com
ze-mail:soji@andrew.cmu.edu
xe-mail:earth@media.teu.ac.jp
{e-mail:mikami@media.teu.ac.jp
xz plane and calculates Lorentz forces acting on the charged particles
at the grid points by simulating the electromagnetic field.The
particle positions are iteratively updated, and the grid size is also
updated after every iteration so that all the particles are always
within range of the grid.
The proposed method simulates charged particles falling through
Earth’s magnetic field. The proposed method determines the type
of the atmospheric particle that collided with a charged particle
based on the colliding position and the composition of upper atmosphere.
The method then calculates the wavelength of the light
radiated from the atmospheric particle based on the generation rate
of the excitation of each wavelength.If the atmospheric particle collides
with another atmospheric particle before emitting the light, it
loses energy and does not emit the light.
The proposed method renders the image of the aurora by projecting
the colliding positions to the screen and calculating the intensities
of emissions.The decrement of the light can be calculated from
the distance between the screen and the colliding position.Gaussian
filter is applied to the light intensities so that the aurora image is
smooth and continuous.
3 Conclusion
This paper has proposed a new modeling method for visual simulation
of aurora.Figure 1 shows snapshots of an animation generated
by the implementation.It successfully visualizes a band of
aurora fluttering.The method can reproduce the aurora movement
such as splitting and rotation and can render the aurora emission in
color. Future work of this research includes improving the modeling
method for simulating the magnetic variations and real-time
processing.These enhancements will make the proposed method
useful for wider variety of visual simulations.
References
G. V. G. BARANOSKI, JON ROKNE, P. S. T. T. R. B. 2003.
“simulating the aurora”. Visual. Comput. Animat, 43–59.
G. V. G. BARANOSKI, J. W. 2005. “simulating the dynamics of
auroral phenomena”. ACM Transactions on Graphics 24, 37–59.
TOMOKAZU ISHIKAWA, YONGHAO YUE, K. I. Y. D. T. N. 2011.
“modeling of aurora borealis using the observed data”. Proc. of
SCCG, 35–38. |
|