Student Projects

Bachelor Projects

200101: Tomographic reconstruction of flames emanating from a matrix burner

A matrix burner, shown in Fig. 1, which constitutes 21 laminar diffusion jet flames should be investigated. This project will include two main tasks:

  1. Perform phantom studies using a numerical field that replicates the flame configuration.
  2. Build up the experiment and obtain the images required for reconstructing the chemiluminescence field using our in-house tomographic algorithm. An example of a highly turbulent flame that was previously reconstructed with the algorithm in shown in Fig. 2.

Multiple projections of the flames are needed from different angles around the burner to perform the reconstructions. Since the flames are steady, a single CCD camera can be used in the experiment, and either the burner is rotated about an axis, or the camera should be rotated around the burner.

For further information please contact:

Matrix Burner
Fig. 1: Matrix burner

Swirl Flame
Fig. 2: Previously reconstructed swirl flame

200102: Tomographic data management in MATLAB

Tomographic experimental data, in the form of chemiluminescence images obtained from different angles around a flame, are used to reconstruct the 3D chemiluminescence field of the flame. Existing MATLAB scripts are currently used to post-process different data types in two main steps:

  1. The experimental images are in tiff file format. These raw data files are post-processed to produce pgm image files which are needed for the tomographic reconstruction algorithm.
  2. The 3D reconstructed field data is generated in h5 file format that is readable by MATLAB. The h5 data files are post-processed to extract various information including cross-sectional and vertical slices at different orientations. Additionally, cross-correlation between different slices are calculated and the data is stored in .dat files. Line plots of pixel intensity data from the slices are also generated.

The candidate is required to organise all the existing MATLAB scripts into one main script, for each of the two main tasks that are given above, which calls all the individual sub-scripts accordingly. For this purpose, the data files and directories in each case also need to be organised appropriately. Further information may need to be extracted from the tomographic volume data such as iso-surfaces and further line plots, for which new sub-scripts need to be written, and which need to be incorporated within the MATLAB scripts. The candidate must have a good grasp of MATLAB and programming experience. Appropriate modularisation and documentation of the resulting code will be a major criterion for the resulting grade.

For further information please contact:

200103: Tomo-hedgehog precision 3D camera mount

Multiple cameras need to be positioned around a field of interest, e.g. a flame, to obtain images from different angles which can be used to reconstruct the field directly in 3D. In this project such a camera mount must be designed and made. The mount must be rigid enough to hold the weight of the cameras (<90 g each) and robust enough to withstand transport to different laboratories. The cameras must not move when fixed on the mount. Most importantly, the location of each camera in world coordinates must be known and hence the mounting system must be designed with precision. A total of 30 cameras are available, dimensions approximately 29 x 29 x 54 mm. The mount must accommodate at least 12 cameras. An example of a possible design is illustrated below.

For further information please contact:

Hedgehog Mount1
Hedgehog Mount2

200104: Design of a filter mount for tomography tests

Filter Mount
Prototype of a modular
design using Lego bricks,
by a previous student
Raimund Huebner

Optical filters are typically mounted in front of the camera lens within the tomography setup, to make measurements of particular wavelengths of light emitted from the test volume. In this project, a suitable and low-cost filter mount system is to be designed and manufactured for use in the multi-camera setup, that contains 30 cameras. The filter mount should allow the bandpass filters to be replaced between tests without disturbing the camera arrangement. Filters sometimes need to be exchanged after the camera arrangement has been calibrated. Following calibration, the cameras must not move or tilt in any direction. Therefore, the filter exchange process should ideally not require any contact to be made with the cameras or the lenses.

​For further information please contact:

Master Projects

200201: 3D Imaging of a spray flame by background-oriented schlieren tomography

Background-Oriented Schlieren (BOS)

BOS is a diagnostic tool that is used to infer the refractive index field, n. Light travels in straight lines when n is constant within a region of interest. However, the light rays will bend if the refractive index changes along their path. A variable n-field can be generated by many sources, for example a flame or supersonic flow. In a BOS experiment, a camera is focused onto a background image that contains a random pattern. A reference image of the background pattern is recorded and compared to an image of the pattern when a variable n-field is placed between the camera and background. Due to bending of the light,  specific points on the background image will exhibit deflections which can be calculated from the image pairs (reference and deflection).

Project description

The aim of this project is to take BOS measurements for a spray burner that is used for nanoparticle synthesis. The 3D refractive index field should be calculated using our in-house tomography algorithm. Prior phantom tests are also necessary to establish the expected theoretical accuracy of the results. The candidate should have programming skills, and any experience in experimental work will be beneficial.

​For further information please contact: or

Bos Setup
The group‘s generic multi-camera BOS setup

200202: 3D Imaging of a spray flame by multi-colour flame tomography

The light emitted by a flame can be measured simultaneously from different angles around the field, using a multi-camera tomography arrangement. In this project, the aim is to simultaneously measure the naturally occurring chemiluminescence that is due to the chemical reactions within the combustion zone of a spray flame that contains multiple streams. In addition, vaporised alkali metals, such as sodium chloride (regular table salt!) will be added to the flow streams. The alkali metals will illuminate in different colours where the gas temperature is high, on the burnt side of the flame (colourful flame). The simultaneous chemiluminescence and alkali metal illumination measurements, which are multi-colour, should be used by our in-house tomography algorithm to reconstruct each field separately. The results should be utilised to analyse the combusting flow zone, and discuss the interaction between various illuminated regions that originate from different streams. The project requires interest and competency in experimental and computational work. The candidate should devise a meaningful way in interpreting the data and quantifying features of interest.

​For further information please contact:

Multi Color
Example of a time-averaged multi-colour
flame reconstruction of a turbulent stratified
flame from a previous project by Cheau Tyan Foo.

200203: Experimental validation of the spatial resolution from tomographic imaging

We obtain instantaneous and time-averaged 3D information from experiments by combining different types of measurements with computed tomography (CT). For example, flame chemiluminescence measurements are used by a CT algorithm to reconstruct the structures within a combusting flow in our computed tomography of chemiluminescence (CTC) techniques. The aim of this project is to design an experiment which allows us to determine the spatial resolution achievable by our CTC methods. solid structures of known sizes, that represent a flame, should be imaged by the array of cameras in the setup for CTC. The structures should be opaque, and emit light in the visible range. The images will then be used to reconstruct the scene and the size of the calculated structures should be compared with the physical size. The tests should be conducted for different types of structures, representing different complexities within the scene and a metric for quantification of the spatial resolution should be devised.

​For further information please contact:

Tomo Setup Project
Image of the existing tomography setup around a burner.

200204: Tomographic particle tracking velocimetry (Tomo-PTV)

Computed tomography (CT) can be combined with different types of measurements, such as emission or absorption, to estimate the volumetric distribution of  different parameters. The measurements are made from different angles around a probe volume (views), and are fed into an algorithm that calculates the three-dimensional (3D) field. In this project, we want to combine PTV with CT using our in-house tomography algorithms, to produce the 3D location of particles in a complex turbulent flow field. A phantom study needs to be performed first, to adapt the tomography algorithm and characterise its behaviour. A phantom is an exactly known field, that can be generated numerically through simulations for example, and allows for a quantitative analysis of the tomography by comparing the original and reconstructed phantoms. The study must look at the effect of particle concentration and number of views on the accuracy of the reconstructed particle field.

The candidate must have strong numerical skills, understanding of fluid mechanics - in particular turbulence. A summary of the thesis shall be turned into a research paper that can be submitted to an international journal or conference.

​For further information please contact:​
PTV concept

200205: Finding the optimal camera arrangement in 3D space for flame tomography

Several cameras are usually arranged around a burner to image the naturally occurring chemiluminescent light from the flame. The images are then used by a tomography algorithm to calculate the 3D chemiluminescence field. The number and arrangement of the cameras affects the quality of the reconstructed fields. A phantom study should be conducted to systematically vary the arrangement of multiple cameras in 3D space around a burner, to determine the best camera distribution. Our in-house evolutionary reconstruction technique can be used for this. The candidate must be very competent in programming skills and be able to adapt the existing code accordingly and work with it.

For further information please contact: or

Cam Location 3d
Illustration of camera location in 3D space,
Unterbeger et al., IEEE International Conference on
Image Processing (2019)


200206: Web-hosted tomographic reconstruction on any device: Go3D APP

The aim in this project is to reconstruct the instantaneous 3D flame shape using multiple images of the chemiluminescence light that are captured from different angles around the flame using a mobile phone or tablet. Our existing calibration and tomography algorithms should be combined in an App for automated online reconstructions. Familiarisation with an App development code, e.g. RStudio’s Shiny is required. The handling of our calibration and tomography procedure should be completed, at first offline. A server account has to be prepared on which the App should be executed with a simple first demonstration case. The candidate must have very high competence in programming, and very good grades in mathematics, computer graphics, numerical courses and other related subjects. Previous experience in App development will be beneficial.

For further information please contact: or

Go 3d App

200207: Camera calibration

For a generic tomographic reconstruction algorithm highly accurate calibration of the cameras' intrinsic and extrinsic parameters (e.g. focal length and location) is necessary. Having sound parameters will help  to decrease various artefacts and improve the reconstruction quality. The student will design a calibration target that is easy to manufacture and handle, but leads to a set of optimal parameters. The influence of the shape of the target on the parameters can be assessed by an existing in-house calibration algorithm. In a first step of the development process the student will become familiar with the code and test different target shapes. In a second step the target, based on the previous findings needs to be manufactured, e.g. with help of a 3D-printer. In the last phase of the project the target will be tested under real world conditions with our tomographic setup.

For further information please contact: or