Real Time Detection, Prediction and Reconstitution of Rain Drops

Authors: R. Burahee, B. Chassinat, T. de Laclos, A. Dépée, A. Sastim

Abstract:

The purpose of this paper is to propose a solution to detect, predict and reconstitute rain drops in real time – during the night – using an embedded material with an infrared camera. To prevent the system from needing too high hardware resources, simple models are considered in a powerful image treatment algorithm reducing considerably calculation time in OpenCV software. Using a smart model – drops will be matched thanks to a process running through two consecutive pictures for implementing a sophisticated tracking system. With this system drops computed trajectory gives information for predicting their future location. Thanks to this technique, treatment part can be reduced. The hardware system composed by a Raspberry Pi is optimized to host efficiently this code for real time execution.

Keywords: reconstitution, prediction, detection, rain drop, real time, raspberry, infrared