DeepFlow

The DeepFlow project (DEEP learning for FLOWing water quality monitoring) will develop an autonomous and intelligent continuous monitoring system for the visual detection of contaminants in water, such as hydrocarbons, in both natural and industrial environments.

The solution is based on computer vision and deep learning technologies for the visual detection of contaminants using cameras placed in strategic positions in channels, rivers, and treatment ponds. These images are processed with Artificial Intelligence (AI)-based algorithms to recognize visual patterns of contamination. In the case of hydrocarbons, the models are trained to recognize the formation of multicolored iridescence or shiny films caused by thin oil films on the surface. Finally, this integrated system will generate automatic notifications in case of anomalies.

With this technology, DeepFlow aims to overcome the limitations of traditional detection systems, which are subjective, have limited coverage, and the use of precise sensors usually has a high cost.

DeepFlow is funded by the European Union – Next Generation EU through the Plan de Recuperación, Transformación y Resiliencia conveyed by the Department of Business and Labor of Generalitat de Catalunya.