基于光谱技术的对虾养殖水质预测系统的研究

Research on Water Quality Prediction System for Shrimp Farming Based on Spectral Technology

随着水产养殖业的快速发展,对虾养殖水质的实时监测与预测显得尤为重要。本研究提出一种基于光谱技术的对虾养殖水质实时预测系统,通过高光谱成像仪获取水质原始光谱数据,结合卷积平滑滤波(Savitzky-Golay Smoothing,SGS)和连续投影算法(Successive projections algorithm,SPA)进行数据处理,并采用最小二乘支持向量机(Least Squares Support Vector Machine,LS-SVM)对对虾养殖水质进行预测。预测模型的决定系数(R2)和均方根误差(RMSE)分别为0.85 mg/L和14.82 mg/L,表明其在对虾养殖水质预测中具备优越的性能。

With the rapid development of aquaculture,real-time monitoring and prediction of water quality in shrimp farming has become particularly important.This study proposes a real-time prediction system for shrimp aquaculture water quality based on spectral technology.The raw spectral data of water quality is obtained through a hyperspectral imager,and the data is processed using Convolutional Smoothing(SGS)and Continuous Projections Algorithm(SPA).The Least Squares Support Vector Machine(LS-SVM)is used to predict the water quality of shrimp aquaculture.The determination coefficient(R 2)and root mean square error(RMSE)of the prediction model are 0.85 and 14.82mg/L,respectively,indicating its superior performance in predicting water quality in shrimp farming.