An intelligent IoT-based control and traceability system to forecast and maintain water quality in freshwater fish farms

摘要

The quality and safety of aquatic products are increasingly important in China. In this study, we developed an Internet of Things (IoT)-based intelligent fish farming and tracking control system that includes a forecasting method that enables automatic water quality management and supports tracking the breeding and selling of freshwater fish. This system can assist fish farmers to intelligently control and manage fishpond water-quality treatment equipment and assist consumers in tracking and viewing historical farming process data using the QR code tag of an aquatic product, which can raise revenue of fish farmers and safeguard the food safety of consumers. We also propose a set of water-quality indicator forecasting methods for a fishpond intelligent management module that first detect and remove abnormal data using the local outlier factor (LOF) algorithm after compared with DBSCAN. Then, the key fishpond data are analysed, modelled and predicted using the model tree algorithm, allowing water-quality indicators to be addressed in advance and maintained within a safe range that complies with standards. The experiments verified that the mean values of the predicted data generated by the M5 model tree algorithm were closer to those of the training data than Cubist, RF, GBM algorithm, and strong correlations were found between the predicted data and the verification data. Moreover, the mean absolute errors of our method are small relative to the data means, indicating that the proposed method can effectively and accurately forecast water-quality indicators.