Geography Reference
In-Depth Information
Fig. 18.9
A scatter plot
generated in our
climate@home project
18.5.2
Multidimensional Rendering and Visualization
Recent decades have generated big data in three dimensional space, time, and mul-
tiple principle component variables. Exploring and analyzing valuable information
hidden in the large volume of data is a difficult task for scientists (Keim
2002
).
Visualizing these data with 3D intuitive effect can help us explore potential patterns
in data, understand complex phenomena and predict the trend for making decision.
With the development of 3D visualization technologies, traditional 2D visu-
alization like map, plot, bar charts, can be complemented with new technology
to visualize and analyze 4D data. There are a large number of 3D visualization
techniques which can be used for visualizing 4D big data. Generally speaking, these
techniques are divided into two categories: 3D isosurface and Volume rendering. 3D
isosurface is a 3D extension of isoline. It is the surface that represents points of a
constant value (e.g., pressure, temperature, density) within a 3D space. Marching
cubes algorithms (Lorensen and Cline
1987
) is a popular method of constructing
such isosurfaces from high resolution 3D datasets by creating triangle shaped
surfaces with the same vertices density as 3D data. Volume rendering is a technique
for visualizing 3D arrays of sampled data (Robert et al.
1988a
,
b
), which is used for
the applications in medical imaging and scientific visualization. The classic direct
volume rendering algorithms are categorized into volume ray casting, splatting,
shear warp and texture mapping. Sometimes, a combination of these techniques
is a better way to solve specific visualization problems.
In many cases, regular visualization algorithms cannot satisfy the requirement for
rendering with intensive computing requirements. To improve efficiency of visu-
alization, many scientists have focused on optimization techniques. Take volume
rendering as example, its use has been limited by its high computational expense.
Search WWH ::
Custom Search