News

Multimodal Imaging for vision through turbid media

UPC-CD6 has been granted project MISTED within the “I+D+i projects” Call from the Science of Innovation ministry of Spain. MISTED (Project PID2020-119484RB-I00) stands for “Multimodal Imaging for viSion through Turbid mEDia” with Prof. Santiago Royo as Main Researcher, and continues the research line on vision thought turbid media which started in the former project of the same call, RESULTANT (FIS2017-89850-R, ImpRoved lasEr Sensing in mUlTipLe scAtterIng media usiNg polarizaTion).

Within MISTED, special emphasis is given to the development of a detailed Monte Carlo algorithm for the simulation of imaging through turbid media. The algorithm will embed non-regular particles in turbid media, extending the simulation of light propagation to sand and smog particles, generalizing the use case of fog, where perfect spherical particles of water are considered. Both lidar imaging and polarization imaging will be analysed from the model and from experimental results, obtained at a large-scale fog chamber. Equivalently, improved data fusion algorithms will be developed to mix different imaging modes, including 3D lidar, polarimetric, RGB, and thermal imaging. Also, deep learning algorithms for perception in turbid media will be tested and developed using different strategies which take advantage of the multimodal nature of the data. Finally, a simple demonstrator of an underwater lidar (at 532nm) will be built and set to work.

The project aligns with Challenge 4 of the Plan Nacional, related to Sustainable, Intelligent and Connected Transportation, as well as to Challenge 9 on Security, Protection and Defence.

Image credits: Markus Spiske on Unsplash