Background

Astronomical data presents certain challenges that affect the approaches taken when developing a visualisation system. Possibly the greatest challenge is how to effectively handle the size of the data that is produced by astronomy recording instruments as high-resolution data contains millions of data points which are worth gigabytes of information. Data processing and visualisation would be done by a remote server which would give a wider range of people access to the data regardless of the computational power of their local computers. It also reduces the need to transfer the data between client computers and would aid astronomers towards more effective data exploration and analysis.

Next

Goals

The goal for this research project was to construct a proof of concept system that implements a strategy for allowing any average computer to produce visualisations of very large data cubes in a lag-free manner. We formulated a series of research questions which we intend to answering through the completion of this project. These questions mainly related to the usability of the new system, the libraries used for compression and rendering as well as the scalability of using large data sets.

Next
Next

Research Questions

  1. How can the data be pre-processed using methods such as mipmapping on the server-side of the system to increase the speed of data processing and transfer?
  2. Does ZFP compression library significantly compress the data to achieve better transfer speed over a network from server to client?
  3. Does VTK graphical rendering library produce the best results in terms of speed and rendering?
  4. How can the client and server approaches be combined to facilitate high-resolution data rendering over a network while maintaining usability?
  5. How does the hybrid model approach scale in terms of rendering time, transfer latency, and time from user interaction and feedback over increasingly larger data sets?
Next

System Overview

Our prototype is using a server-client architecture similar to CARTA. The hybrid approach implies the combination of the server and the client in a manner that eliminates the drawbacks and limitations of a purely one-sided approach.

Next

Conclusions

The hybrid approach system brings forward the advantage of using any commodity machine as client over a reasonable network to interact with 3D astronomical data in almost real-time. It enables 3D visualisation of astronomical images worth gigabytes or even terabytes of data. We believe that this system is a right way of proceeding towards better 3D visualisation in the future when more detailed images with less noise will be collected by modern telescopes.