MSc Thesis subject: Can a modified digital camera be better than a commercial multispectral sensor?An open-source multispectral system: from modified camera to image-product.

Most (if not all) commercial multispectral sensors are based on silicon chips, which are present on your cell phone camera and also on Landsat.
If you are an amateur photographer or camera enthusiast, you will know that, there are a number of key-points that make a good camera.
The goal of this thesis is to enable the candidate to understand and develop a 3-band sensor that could match or surpass the quality of a commercial multi-spectral system.
Sequentially, the candidate will employ such camera on an aerial platform and develop a classification or regression model for a case-study of his own preference.

Many phenomena in vegetation can be studied through the analysis of the correct spectra (range of wavelengths) and radiometric response (intensity of light). A regular camera measures both colours and its intensity variation. The spectra which a camera captures can be constrained using a filter.

Yet, if you already took pictures of the same object with different cameras, you should know that the output (intensity of each colour) can vary dramatically.
Among other factors, that is due to the camera spectral sensitivity and dynamic range.
This thesis focuses, then, on the development of a spectral and radiometric standardization method for modified digital cameras; allowing cross-comparison between different sensors and in different light conditions.
Additionally, the imagery can be processed in an open-source platform; thus, providing low-cost, quality assessed multispectral data.


  • Compare and contrast a modified digital camera and a commercial multispectral sensor
  • Develop a method where a filter-based modification could provide the same output in any digital camera model.
  • Apply the modified digital camera in a study-case of the candidate’s choice.



  • A hacker/maker approach to problem solving.
  • Interest on multispectral data/drone for vegetation analysis
  • Full understanding of the Han et al (2012) paper listed above.

Theme(s): Sensing & measuring