Skip to main content

Visualisation of 3d datasets

With the advances of computer technology, computational MHD models have become more prestigious. Three dimensional datasets have become common place. But since presentation techniques are limited to only two dimensions (e.g. overhead, slides, computer screen), the sense of depth is lost. There are many options put forward to overcome this: shading, reference lines, light sources. Researchers at the Solar Theory Group of the University of St Andrews put forward anaglyphing as a way of simulating depth. This method uses special glasses to decode the picture.

From my background of being an amateur astronomer, helping out at the local planetarium in Genk, Belgium, I came across chromostereoscopic glasses and I saw the potential for solar physics. A web page has been designed to highlight the usefulness of chomostereoscopy and also anaglyphing for 3d datasets in solar physics:

St Andrews Solar Theory Group Visualization website

Chromo-stereoscopy is a technique of simulating depth perception in two dimensional images. Three dimensional datasets are projected onto two dimensions, leaving the third, normal dimension to be used as a scaling for assigning colour to the pixels representing the projected data points. The depth perception is coded into the image with the use of colour, hence the name chromo-stereoscopy.


The 3d effect is created by looking at the picture through special double prism refraction ChromaDepth TMC3D glasses which decode the image and show the depth. These glasses are, unlike the anaglyph glasses, clear. The trick is the incorporated micro-optics. In principle the glasses are two back to back prisms pointing in opposite directions. The first prism is a high dispersive prism, decomposing the light. The second prism is a low dispersive prism which counter-acts the angle deviation caused by the first prism.

Chromo-stereoscopy is very well suited for visualizing 3d magnetic field structures

Three-dimensional magnetic fieldline structures are very well suited for the technique chromo-stereoscopy because there is a strong interaction with the background in the image. The fieldlines should be drawn with a certain thickness to show the colour well.
When studying three-dimensional magnetic field structures, it is sometimes very difficult to imagine how it looks like from 2d flat projections. With chromo-stereoscopy you really see the 3d structure.

Potential magnetic field reconstruction from NSO/GSFC synoptic chart magnetogram CR1724
(Aad van Ballegooijen and Duncan Mackay)

Seperatrix surface of a potential-field model for a filament channel.
Data courtesy of Duncan Mackay, University of St Andrews


Chromo-stereoscopy and iso-surfaces

Chromo-stereoscopic magnetic fieldline structures combined with iso-surfaces is more difficult to show the 3d structure of iso-surfaces using chromo-stereoscopy. This is because the background colour is not close-by the structures. This can be overcome with putting lines (in the background colour) on the iso-surface which follow the curvature.

1. The whole image is produced with a shading routine. The magnetic fieldlines are shown as thin tubes.

2. The iso-surfaces have been shaded using a light-source. The magnetic fieldlines are still rendered in the chromo-stereoscopic colour code.

3. Both magnetic fieldlines and iso-surfaces are rendered in the chromo-stereoscopic colour code.

4. The background colour on top of the iso-surfaces appears at the depth of the colour, and helps enhance the shape of the iso-surfaces.

Formation of Prominences by Flux Convergence
Images courtesy of Klaus Galsgaard, University of St. Andrews


Basic procedure

The basics of producing chromo-stereoscopic pictures if you have a 3d dataset is very easy and quickly achieved using packages like IDL (Interactive Data Language). The steps to do are:

  • Load in a correct colour table which is linear in depth (e.g. the ChromaDepth TMC3D RGB colour table).
  • Rotate 3d dataset according to the view you want (from (x,y,z) --> (x',y',z')), and use two coordinates (say x' and y') for projection of the image and one coordinate (say z') as a linear scaling for the colour at every datapoint (colour = blue + z'/dz (red - blue) ).
  • Do a plot of the data with color coding.

You can download an IDL program I wrote and play around with it, alter it to your liking. I'm certain there are better IDL programmers out there than me.

To download the routine: depth3D.pro
You will also need the colour table: RGBpalette



Publication

Download a paper, published in Solar Physics:

Verwichte,E. and Galsgaard,K.
On the visualization of 3d datasets
1998, Solar Physics, 183, 445-448.

Download a postscript copy

Abstract:

The effective visualization of three dimensional (3d) datasets, both observationally and computationally sourced, is becoming common in solar physics. We present example plots of data from a 3d magnetohydrodynamical simulation, where depth perception is simulated using chromo-stereoscopy. The depth information is coded into the images using colours. When such images are viewed with double prism refraction ChromaDepth TM 3D glasses, a pronounced 3d effect is achieved. This visualization method is especially suited for working with and presenting computationally derived 3d datasets.