| n order to monitor the quality of paddy field irrigation water real-time and fastly, and ensure thesafety of food production, a method based on machine vision for detection of copper ion concentration in the paddyfield irrigation water was proposed in this paper. Detection tests based on machine vision of different concentrationsof copper ion solution was designed using detection tester with machine vision. By extracting the eigenvalues of testimages, methods including index regression, logistic regression, two order polynomial regression and linear regressionwere used to build prediction model of copper ion concentration, which reduced the influence of human cognitive ontesting results. The results showed that, in the range of 0-300 mg/L, the prediction model of copper ion concentrationwas achieved with correlation coefficient of 0.9438 and root mean square error of 19.9563 mg/L for training set andcorrelation coefficient of 0.9191 and root mean square error of 9.7889 mg/L for prediction set. The relative error waskept below 8 percent, basically achieving the rapid detection of paddy field irrigation water in real time. The resultsof this study provide the basis for the further development of paddy field irrigation equipment utility. |
