An IoT framework for quality analysis of aquatic water data using time-series convolutional neural network

Area of Science:

• Aquaculture
• Environmental Monitoring
• Artificial Intelligence

Background:

• Effective water quality monitoring is crucial for sustainable aquaculture.
• Traditional water quality analysis methods present significant challenges in fish farms.
• The need for advanced, automated solutions is pressing.

Purpose of the Study:

• To develop and evaluate an Internet of Things (IoT)-based deep learning model for real-time fish farm water quality monitoring.
• To enhance the accuracy and efficiency of water quality analysis compared to existing methods.
• To utilize a time-series convolution neural network (TMS-CNN) for spatial-temporal data analysis.

Main Methods:

• An IoT-based system was integrated with a deep learning model, specifically a time-series convolution neural network (TMS-CNN).
• The TMS-CNN model was designed to effectively process spatial-temporal water quality data, capturing complex dependencies.
• Water Quality Index (WQI) was calculated using correlation analysis, followed by class label assignment and time-series data analysis.

Main Results:

• The proposed TMS-CNN model demonstrated high accuracy (96.2%) in analyzing water quality parameters relevant to fish growth and mortality.
• The model's performance surpassed the current best model (MANN) by achieving a higher accuracy rate (91%).
• The TMS-CNN effectively handled spatial-temporal dependencies in water quality data.

Conclusions:

• The IoT-based TMS-CNN model offers a highly accurate and effective solution for fish farm water quality monitoring.
• This deep learning approach significantly improves upon traditional methods and existing models.
• The findings support the adoption of advanced AI techniques for optimizing aquaculture management and fish health.