Based on floc morphology parameters, a partial least squares model was established to solve the problem of real time measurement of COD, BOD5, TN and TP during the operation of sewage treatment. The floc morphological characteristics of sludge were extracted by image analysis technology. Based on the correlation, the input variables of the model were selected from the extracted morphological parameters and the operation parameters during the experiment. The prediction model of the relationship between input variables and four water quality indexes was established by partial least squares. 〖JP2〗The results indicated that the cross validity coefficients of COD, BOD5, TN and TP were 0.736, 0.682, 0.839 and 0.618, respectively. There was a significant correlation between 〖JP3〗floc morphology parameters and water quality indexes. This model could be used to predict effluent quality. |