Design of Enterprise Wastewater Supervision System Based on Federal Learning

This paper proposes a federated learning-based system for enterprise wastewater monitoring system, using image pre-processing, principal component analysis method (PCA), optimising BP neural network centred on improved adaptive genetic algorithms to achieve the refinement of wastewater detection and ensure low-cost operation while treating the water quality up to the standard. In order to solve the problem of sharing the data while also protecting the enterprise's privacy, the platform adopts federated learning to build trust boundaries, exploring the new road of intelligent enterprise sewage supervision system.