A novel determination method of milk somatic cell count based on dielectric relaxation parameters

Abstract

Rapid and accurate raw milk somatic cell count (SCC) monitoring is vital to identify cow mastitis and dairy quality evaluation. This study aimed to explore the influence mechanism of milk SCC on dielectric spectra and the potential of dielectric relaxation parameter (DRP) to predict milk SCC. The dielectric spectra of 272 raw milk samples with different SCC at 20–4500 MHz were measured, and the effects of milk SCC on the DRPs were analyzed. The models based on full spectra (FS), characteristic variables extraction, and DRPs methods were established to predict milk SCC. The results showed that the Cole-Cole plots exhibited apparent differences between sample groups with different SCC. The DRPs-NNuSVR model had the best prediction performance, with the standard error of prediction of 0.166 log SCC/mL and residual predictive deviation of 2.538, better than the FS and characteristic variables extraction models. This study explores the influence of milk SCC on dielectric relaxation and proposes a novel milk SCC prediction method, which helps in dairy farming and raw milk quality assessment.