Computational optics is a basic discipline for research in optics and for the development and optimization of applications in optics, like lasers, lenses, optical nanostructures or imaging. Therefore, there are several connections of the main topic “computational optics” to the other main topics. The research and training activities in computational optics can be divided into the following research directions (modules):

  • Imaging lab and pattern recognition
  • Simulation and optimization of optical systems
  • Simulation of optical waves
  • Laser simulation

 Imaging lab and pattern recognition.  Image processing, computer vision and pattern recognition research is tackling the hard problem of designing efficient algorithms for the analysis of image data acquired by a wide range of optical systems. Nowadays the used sensor technology ranges form standard CCD cameras over X-ray systems to optical coherence tomography devices. For the development of cognitive visual systems it is mandatory to provide core technologies of computational optics. For the quantitative analysis of image data, the mathematical modelling of the transfer function of the optical system and the calibration of the optical system are crucial. This includes, for instance, the projection mapping from 3-D to 2-D, lens distortion or the attenuation of x-ray quanta while being propagated through material. For the reconstruction of higher dimensional image information, numerical algorithms have to be developed that solve the ill-conditioned inverse problems robustly and efficiently. Examples are the 3-D reconstruction of volumes from 2-D x-ray projections, the computation of 3-D surface data using optical sensors like stereo cameras or the computation of in-vivo data based on phase modulated light as it is done in time of flight imaging. The major focus of current research is on the development of the technologies for the mathematical modelling and algorithmic description of optical systems used for image analysis. A huge potential for innovation in this emerging field is seen in the close interdisciplinary collaboration of experts in physics, engineering, computer science and medicine and the resulting synergies.

Simulation and optimization of optical systems. There are a variety of optical simulation methods which are adapted to a specific problem. But, in many cases several simulation methods can be combined if a proper interface between the different physical models describing the light distribution can be found.

One example is the combination of a rigorous diffraction theory for periodic structures and ray tracing. By doing this, also diffractive optical elements with very short period sizes (of only some wavelengths) can be simulated together with a complete optical system consisting of lenses, mirrors and so on. Of course, such a combination of different methods is only valid approximately. For the combination of a rigorous diffraction theory and ray tracing it is for example assumed that there is a strictly periodic and infinitely extended grating, whereas in practice a diffractive optical element has a locally varying period size and a finite size of the element. This approximation is valid with good accuracy in most practical cases


Simulation of optical waves. Research work is done exemplarily in thin film solar cells.  Thin film solar cells are an innovative low cost technology in renewable energies. To increase absorbing of light in solar cells, light trapping is a very importing research topic in photovoltaics. The main concepts to improve light trapping are suitable nanostructures, rough interfaces between the different layers and use of plasmonic effects to increase the intensity of light in thin film solar cells. The aim of research is to perform simulations on high performance computers to improve light trapping in thin film solar cells and to improve simulation techniques for analyzing optical waves in thin film solar cells.


Laser simulation. Solid state lasers are widely used lasers for industrial and medical applications. In particular diode-pumped solid state lasers with amplifiers are used to obtain Q-switch lasers with very high pulse energy. The aim of research is to improve the simulation techniques for such kind of lasers by using Finite Element discretizations for Maxwell’s equations. By this approach an accurate simulation of the optical wave inside and outside the resonator is possible. This includes losses caused by apertures and thermal induced stress birefringence.