Augmented reality is a relatively new domain and is emerging as an alternative so-lution to virtual reality. In a givenvirtual realityapplication, the user is emerged in a completely synthetic environment. In contrast to this,augmented realitydoes not completely suppress the real environment, the real world is augmented by relevant additional information to the user by means of a computer [Bimber and Raskar 2005]. Indeed, augmented reality is situated between virtual reality and the real world and uses the advantages of both. The perception of the real world is augmented with information which the human senses could not perceive. For a comprehensive review of the field, the reader should refer to the field overview pa-per by Azuma [Azuma 1997], or to the recent book by Bimber and Raskar [Bimber and Raskar 2005].
There are several ways of building an augmented reality system. The widest spread systems use head mounted displays, though recently hand held and spa-tial displays are also emerging [Bimber and Raskar 2005]. Generally, almost all types of displays used in augmented reality can be classified intovideo see trough displays and optical see trough displays . A video see through system is one where the user only sees the environment displayed onto a screen as recorded by an attached video camera. It has the disadvantage that the user is less likely to perceive the scene as realistic (a similar experience is using digital cameras with electronic viewfinder), but the advantage is that overlaying additional information and system calibration is much easier. In contrast to video see through systems, optical see through systems overlay all additional information directly onto the image of the real world perceived by the user. This is achieved by means of semi-transparent displays or beam splitters. Optical see through systems have the advantage of a higher realism, but are more complicated to align and calibrate.
The augmented reality system proposed by us as an educational aid in astronomy uses an optical see through system and is presented in Chapter 8. Augmented
2.6 Augmented Reality 19
Figure 2.7: A typical augmented reality application for architectural planning.
Image courtesy of the VR Center for the Built Environment [VR Centre for the Built Environment 2004]
reality systems have a very large domain of applicability nowadays. We men-tion just some of the fields where this technology is applied: industrial, medicine, architecture and construction, entertainment, military and education applications.
20 Chapter 2: Background
Chapter 3 Related Work
In this chapter we review related work to the solutions presented in this thesis.
There is a great amount of work in the field of rendering and reconstruction of astronomical objects in the computer graphics as well as in the astrophysical lit-erature. We also present here augmented reality applications and educational as-tronomy devices related to the augmented astronomical telescope proposed by us.
3.1 Rendering of Atmospheric Phenomena
Realistic renderings of atmospheric phenomena has been a well researched topic in the computer graphics community. We review papers focusing on solar disc rendering, as well as work on rendering the sky.
Some publications focus on physically correct and realistic atmosphere simula-tions, for daytime [Preetham et al. 1999], [Nishita et al. 1996] as well as for night-time [Wann Jensen et al. 2001]. Also, a system for rendering the atmosphere from a viewpoint situated in space is presented by Nishita et al. [Nishita et al. 1993].
These approaches concentrate on fast rendering of the atmosphere, approximat-ing the physics of atmospheric light transport. An evaluation and validation with real world measurements of the Preetham sky model [Preetham et al. 1999] was recently presented by Zotti et al. [Zotti et al. 2007].
A work focusing on a physically correct simulation of the atmosphere during twi-light phenomena is presented in [Haber et al. 2005]. The authors create a model of the Earth’s atmosphere including air molecules, aerosols and water particles
22 Chapter 3: Related Work
and account for Rayleigh as well as for Mie scattering by simulating full radi-ation transfer over a discretized hemisphere. By considering different climatic conditions, corresponding hemispherical twilight skies can be computed.
A different method focusing on the rendering of the solar disc is presented by Bru-ton [BruBru-ton 1996]. In his thesis, ray-tracing through the atmosphere is performed using Lehn’s model [Lehn 1985], and solar disc appearance is simulated from diverse input temperature profiles. However, this work considers only Rayleigh scattering due to air molecules. Thus, different types of sunsets depending on current aerosol distribution in the atmosphere cannot be simulated. Another work based on Bruton’s model concentrates more only on reproducing the correct shape of the solar disc is presented by Sloup [Sloup 2003].
One possible approach is to simulate the green flash or other atmospheric phenom-ena using an approach based on photon mapping [Gutierrez et al. 2004,Seron et al.
2004,Guitierrez et al. 2005]. These papers focus on implementing a “Curved Pho-ton Mapping” algorithm. Although the obtained results are highly realistic, these approaches are slow (due to the photon mapping algorithm), and they lacks any dependence on climate conditions.
A paper describing the theoretic concepts of non-linear ray-tracing was presented by Gr¨oller [Gr¨oller 1995]. He describes data structures and ray object intersection algorithms specially suited to the case of non-linear rays. An in depth review of work related to rendering of atmospheric phenomena is presented by Sloup [Sloup 2002]. As already mentioned in the previous Chapter, for a book describing phe-nomena relating to atmospheric optics the reader should refer to [Minnaert 1954].
In contrast to the already reviewed works, our approach presented in Section 4 concentrates on combining the simplified parabolic model for ray-tracing in the atmosphere presented by Lehn [Lehn 1985] and the climate dependent stratified atmosphere model presented in [Haber et al. 2005] in order to create a ray-tracing system which realistically reproduces several possible sequences of the solar disc at sunset or sunrise.