Robustness of hyperspectral imaging and PLSR model predictions of intramuscular fat in lamb M. longissimus lumborum across several flocks and years

The percentage of intramuscular fat content of lamb meat is a key index of consumer acceptability. Hyperspectral imaging is a potential technique for in-line measurements of intramuscular fat in fresh meat. However, little work has been conducted to investigate the robustness of hyperspectral imaging data and associated multivariate models over time. Fifteen trials consisting of eight independent flocks across five years were used to quantify robustness of partial least squares regression (PLSR) models developed using data collected with the same imaging system. Two models were developed; one using data from the first year of the trials, and a progressive model that cumulatively includes data in chronological order. The two models performed similarly, in terms of the coefficient of determination (R2), standard error of prediction (SEP) and bias, when experimental conditions were consistent. However, under varying imaging conditions, the progressive model was able to account for this variability resulting in higher R2 and lower SEP.

Automated summary

The study found that while the two models performed similarly under consistent experimental conditions, the progressive model was better able to account for variability under different imaging conditions, resulting in higher coefficient of determination (R2) and lower standard error of prediction (SEP).