An IoT Based Water Quality Classification Framework for Aqua-Ponds Through Water and Environmental Variables Using CGTFN Model.

Maintaining water quality in aquatic habitats is critical for the health of aquatic species, particularly fish. This study pioneers an innovative method to water quality classification, leveraging IoT-driven data acquisition and meticulous data labelling with the Aqua-Enviro Index (AEI) by considering the fish habitats. Existing mechanisms fail to capture complex temporal dynamics and depend largely on large amounts of labelled data, exposing fundamental limits. In response, we describe the Deep learning based Convolutional Gated Recurrent Unit Tempo Fusion Network (CGTFN) model, which represents a considerable development in the evaluation of water quality. The model addresses these restrictions by seamlessly merging Convolutional Neural Networks (CNNs) for spatial pattern recognition and Gated Recurrent Units (GRUs) for temporal interactions. The Tempo Fusion mechanism combines spatial, temporal, and contextual data harmoniously, allowing for more sophisticated classifications by recognizing subtle interdependencies among environmental elements. The pioneering CGTFN model outperforms previous models, achieving 99.71 and 99.81% accuracy on both public-env and real-time-env datasets, respectively, exceeding established models at 98.2%. These remarkable findings highlight CGTFN’s disruptive potential in water quality evaluation, bridging the gap between technology and environmental management, with ramifications ranging from aquaculture to resource sustainability.Highlights: In this paper we are working with IoT enabled deep learning based CGFTN model for water quality classification considering impact of environmental variables. The data collected through IoT is labelled using an innovative AEI-driven approach specifically designed for fish habitats. CGTFN model integrates CNNs and GRUs, tackling temporal dynamics limitations. Introduce sophisticated Tempo Fusion, harmonizing spatial, temporal, and contextual data. CGTFN demonstrates disruptive potential, achieving superior accuracy (99.71 and 99.81%) on public and real-time datasets. [ABSTRACT FROM AUTHOR]